Merging r243642, r243643, and r243644:
------------------------------------------------------------------------
r243642 | uweigand | 2015-07-30 07:08:36 -0700 (Thu, 30 Jul 2015) | 29 lines

Add support for System z vector language extensions

The z13 vector facility has an associated language extension,
closely modeled on AltiVec/VSX.  The main differences are:

- vector long, vector float and vector pixel are not supported

- vector long long and vector double are supported (like VSX)

- comparison operators return a vector rather than a scalar integer

- shift operators behave like the OpenCL shift operators

- vector bool is only supported as argument to certain operators;
  some operators allow mixing a bool with a non-bool vector 

This patch adds clang support for the extension.  It is closely modelled
on the AltiVec support.  Similarly to the -faltivec option, there's a
new -fzvector option to enable the extensions (as well as an -mzvector
alias for compatibility with GCC).  There's also a separate LangOpt.

The extension as implemented here is intended to be compatible with
the -mzvector extension recently implemented by GCC.

Based on a patch by Richard Sandiford.

Differential Revision: http://reviews.llvm.org/D11001
------------------------------------------------------------------------

------------------------------------------------------------------------
r243643 | uweigand | 2015-07-30 07:10:43 -0700 (Thu, 30 Jul 2015) | 18 lines

[SystemZ] Add support for vecintrin.h vector built-in functions

This patch adds support for the System Z vector built-in functions.
The API-defined header file has the name vecintrin.h.

The user-level functions are defined in the same style as the clang
version of altivec.h, making heavy use of the __overloadable__ and
__always_inline__ attributes.  Where possible the functions expand to
generic operations rather than specific built-in functions, in the hope
that that form can be optimised better.

Where a built-in routine is specified to require an immediate integer
argument, the __enable_if__ attribute is used to verify the argument is
in fact constant and in the appropriate range.

Based on a patch by Richard Sandiford.
------------------------------------------------------------------------

------------------------------------------------------------------------
r243644 | uweigand | 2015-07-30 08:53:58 -0700 (Thu, 30 Jul 2015) | 21 lines

Fix sanitizer fallout from r243642

The memory-sanitizer build bot reported:

==5574== WARNING: MemorySanitizer: use-of-uninitialized-value
    #0 0x7f03089e15f6 in
clang::Parser::TryAltiVecTokenOutOfLine(clang::DeclSpec&,
clang::SourceLocation, char const*&, unsigned int&, bool&)
/mnt/b/sanitizer-buildbot3/sanitizer-x86_64-linux-fast/build/llvm/tools/clang/lib/Parse/ParseDecl.cpp:6290:11

This is because the "Ident_pixel" variable was uninitialized
in the getLangOpts().ZVector case, but we'd still call into
clang::Parser::TryAltiVecTokenOutOfLine, which uses the variable.

The simplest fix for this without sprinkling !getLangOpts().ZVector
checks all over the code seems to be to just initialize the variable
to nullptr; this will then do the right thing on ZVector.

Checked in to unbreak the build bots.
------------------------------------------------------------------------

llvm-svn: 243699
diff --git a/clang/include/clang/Basic/DiagnosticParseKinds.td b/clang/include/clang/Basic/DiagnosticParseKinds.td
index 1364b98..e4f8599 100644
--- a/clang/include/clang/Basic/DiagnosticParseKinds.td
+++ b/clang/include/clang/Basic/DiagnosticParseKinds.td
@@ -358,6 +358,10 @@
   "'%0' declaration specifier not allowed here">;
 def err_invalid_vector_bool_decl_spec : Error<
   "cannot use '%0' with '__vector bool'">;
+def err_invalid_vector_long_decl_spec : Error<
+  "cannot use 'long' with '__vector'">;
+def err_invalid_vector_float_decl_spec : Error<
+  "cannot use 'float' with '__vector'">;
 def err_invalid_vector_double_decl_spec : Error <
   "use of 'double' with '__vector' requires VSX support to be enabled "
   "(available on POWER7 or later)">;
diff --git a/clang/include/clang/Basic/LangOptions.def b/clang/include/clang/Basic/LangOptions.def
index 8d606a1..c184df7 100644
--- a/clang/include/clang/Basic/LangOptions.def
+++ b/clang/include/clang/Basic/LangOptions.def
@@ -104,6 +104,7 @@
 LANGOPT(ConstStrings      , 1, 0, "const-qualified string support")
 LANGOPT(LaxVectorConversions , 1, 1, "lax vector conversions")
 LANGOPT(AltiVec           , 1, 0, "AltiVec-style vector initializers")
+LANGOPT(ZVector           , 1, 0, "System z vector extensions")
 LANGOPT(Exceptions        , 1, 0, "exception handling")
 LANGOPT(ObjCExceptions    , 1, 0, "Objective-C exceptions")
 LANGOPT(CXXExceptions     , 1, 0, "C++ exceptions")
diff --git a/clang/include/clang/Basic/TokenKinds.def b/clang/include/clang/Basic/TokenKinds.def
index 7a91c9f..8333a4c 100644
--- a/clang/include/clang/Basic/TokenKinds.def
+++ b/clang/include/clang/Basic/TokenKinds.def
@@ -239,6 +239,8 @@
 //   KEYOPENCL  - This is a keyword in OpenCL
 //   KEYNOOPENCL  - This is a keyword that is not supported in OpenCL
 //   KEYALTIVEC - This is a keyword in AltiVec
+//   KEYZVECTOR - This is a keyword for the System z vector extensions,
+//                which are heavily based on AltiVec
 //   KEYBORLAND - This is a keyword if Borland extensions are enabled
 //   BOOLSUPPORT - This is a keyword if 'bool' is a built-in type
 //   HALFSUPPORT - This is a keyword if 'half' is a built-in type
@@ -501,7 +503,7 @@
 ALIAS("read_write", __read_write    , KEYOPENCL)
 // OpenCL builtins
 KEYWORD(__builtin_astype            , KEYOPENCL)
-KEYWORD(vec_step                    , KEYOPENCL|KEYALTIVEC)
+KEYWORD(vec_step                    , KEYOPENCL|KEYALTIVEC|KEYZVECTOR)
 
 // OpenMP Type Traits
 KEYWORD(__builtin_omp_required_simd_align, KEYALL)
@@ -510,9 +512,9 @@
 KEYWORD(__pascal                    , KEYALL)
 
 // Altivec Extension.
-KEYWORD(__vector                    , KEYALTIVEC)
+KEYWORD(__vector                    , KEYALTIVEC|KEYZVECTOR)
 KEYWORD(__pixel                     , KEYALTIVEC)
-KEYWORD(__bool                      , KEYALTIVEC)
+KEYWORD(__bool                      , KEYALTIVEC|KEYZVECTOR)
 
 // ARM NEON extensions.
 ALIAS("__fp16", half                , KEYALL)
diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 6e5dbf2..9d3e2cf 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -1351,6 +1351,13 @@
 def mvx : Flag<["-"], "mvx">, Group<m_Group>;
 def mno_vx : Flag<["-"], "mno-vx">, Group<m_Group>;
 
+def fzvector : Flag<["-"], "fzvector">, Group<f_Group>, Flags<[CC1Option]>,
+  HelpText<"Enable System z vector language extension">;
+def fno_zvector : Flag<["-"], "fno-zvector">, Group<f_Group>,
+  Flags<[CC1Option]>;
+def mzvector : Flag<["-"], "mzvector">, Alias<fzvector>;
+def mno_zvector : Flag<["-"], "mno-zvector">, Alias<fno_zvector>;
+
 def mno_warn_nonportable_cfstrings : Flag<["-"], "mno-warn-nonportable-cfstrings">, Group<m_Group>;
 def mno_omit_leaf_frame_pointer : Flag<["-"], "mno-omit-leaf-frame-pointer">, Group<m_Group>;
 def momit_leaf_frame_pointer : Flag<["-"], "momit-leaf-frame-pointer">, Group<m_Group>,
diff --git a/clang/include/clang/Parse/Parser.h b/clang/include/clang/Parse/Parser.h
index fb9eb8f..8719555 100644
--- a/clang/include/clang/Parse/Parser.h
+++ b/clang/include/clang/Parse/Parser.h
@@ -108,12 +108,13 @@
   /// Ident_super - IdentifierInfo for "super", to support fast
   /// comparison.
   IdentifierInfo *Ident_super;
-  /// Ident_vector, Ident_pixel, Ident_bool - cached IdentifierInfo's
-  /// for "vector", "pixel", and "bool" fast comparison.  Only present
-  /// if AltiVec enabled.
+  /// Ident_vector, Ident_bool - cached IdentifierInfos for "vector" and
+  /// "bool" fast comparison.  Only present if AltiVec or ZVector are enabled.
   IdentifierInfo *Ident_vector;
-  IdentifierInfo *Ident_pixel;
   IdentifierInfo *Ident_bool;
+  /// Ident_pixel - cached IdentifierInfos for "pixel" fast comparison.
+  /// Only present if AltiVec enabled.
+  IdentifierInfo *Ident_pixel;
 
   /// Objective-C contextual keywords.
   mutable IdentifierInfo *Ident_instancetype;
@@ -605,10 +606,12 @@
   bool TryAltiVecToken(DeclSpec &DS, SourceLocation Loc,
                        const char *&PrevSpec, unsigned &DiagID,
                        bool &isInvalid) {
-    if (!getLangOpts().AltiVec ||
-        (Tok.getIdentifierInfo() != Ident_vector &&
-         Tok.getIdentifierInfo() != Ident_pixel &&
-         Tok.getIdentifierInfo() != Ident_bool))
+    if (!getLangOpts().AltiVec && !getLangOpts().ZVector)
+      return false;
+
+    if (Tok.getIdentifierInfo() != Ident_vector &&
+        Tok.getIdentifierInfo() != Ident_bool &&
+        (!getLangOpts().AltiVec || Tok.getIdentifierInfo() != Ident_pixel))
       return false;
 
     return TryAltiVecTokenOutOfLine(DS, Loc, PrevSpec, DiagID, isInvalid);
@@ -618,7 +621,7 @@
   /// identifier token, replacing it with the non-context-sensitive __vector.
   /// This returns true if the token was replaced.
   bool TryAltiVecVectorToken() {
-    if (!getLangOpts().AltiVec ||
+    if ((!getLangOpts().AltiVec && !getLangOpts().ZVector) ||
         Tok.getIdentifierInfo() != Ident_vector) return false;
     return TryAltiVecVectorTokenOutOfLine();
   }
diff --git a/clang/include/clang/Sema/Sema.h b/clang/include/clang/Sema/Sema.h
index db7b6f9..ecbcd94 100644
--- a/clang/include/clang/Sema/Sema.h
+++ b/clang/include/clang/Sema/Sema.h
@@ -8363,7 +8363,8 @@
 
   /// type checking for vector binary operators.
   QualType CheckVectorOperands(ExprResult &LHS, ExprResult &RHS,
-                               SourceLocation Loc, bool IsCompAssign);
+                               SourceLocation Loc, bool IsCompAssign,
+                               bool AllowBothBool, bool AllowBoolConversion);
   QualType GetSignedVectorType(QualType V);
   QualType CheckVectorCompareOperands(ExprResult &LHS, ExprResult &RHS,
                                       SourceLocation Loc, bool isRelational);
diff --git a/clang/lib/Basic/IdentifierTable.cpp b/clang/lib/Basic/IdentifierTable.cpp
index dcb7603..7705834 100644
--- a/clang/lib/Basic/IdentifierTable.cpp
+++ b/clang/lib/Basic/IdentifierTable.cpp
@@ -110,7 +110,8 @@
     HALFSUPPORT = 0x08000,
     KEYCONCEPTS = 0x10000,
     KEYOBJC2    = 0x20000,
-    KEYALL = (0x3ffff & ~KEYNOMS18 &
+    KEYZVECTOR  = 0x40000,
+    KEYALL = (0x7ffff & ~KEYNOMS18 &
               ~KEYNOOPENCL) // KEYNOMS18 and KEYNOOPENCL are used to exclude.
   };
 
diff --git a/clang/lib/Basic/Module.cpp b/clang/lib/Basic/Module.cpp
index 3846fec..4314b41 100644
--- a/clang/lib/Basic/Module.cpp
+++ b/clang/lib/Basic/Module.cpp
@@ -67,6 +67,7 @@
                         .Case("objc_arc", LangOpts.ObjCAutoRefCount)
                         .Case("opencl", LangOpts.OpenCL)
                         .Case("tls", Target.isTLSSupported())
+                        .Case("zvector", LangOpts.ZVector)
                         .Default(Target.hasFeature(Feature));
   if (!HasFeature)
     HasFeature = std::find(LangOpts.ModuleFeatures.begin(),
diff --git a/clang/lib/Basic/Targets.cpp b/clang/lib/Basic/Targets.cpp
index dfe6593..f965ab3 100644
--- a/clang/lib/Basic/Targets.cpp
+++ b/clang/lib/Basic/Targets.cpp
@@ -5725,6 +5725,8 @@
     Builder.defineMacro("__LONG_DOUBLE_128__");
     if (HasTransactionalExecution)
       Builder.defineMacro("__HTM__");
+    if (Opts.ZVector)
+      Builder.defineMacro("__VEC__", "10301");
   }
   void getTargetBuiltins(const Builtin::Info *&Records,
                          unsigned &NumRecords) const override {
diff --git a/clang/lib/Driver/Tools.cpp b/clang/lib/Driver/Tools.cpp
index d46f467..19ebbb9 100644
--- a/clang/lib/Driver/Tools.cpp
+++ b/clang/lib/Driver/Tools.cpp
@@ -4049,9 +4049,11 @@
   Args.AddLastArg(CmdArgs, options::OPT_fstandalone_debug);
   Args.AddLastArg(CmdArgs, options::OPT_fno_standalone_debug);
   Args.AddLastArg(CmdArgs, options::OPT_fno_operator_names);
-  // AltiVec language extensions aren't relevant for assembling.
-  if (!isa<PreprocessJobAction>(JA) || Output.getType() != types::TY_PP_Asm)
+  // AltiVec-like language extensions aren't relevant for assembling.
+  if (!isa<PreprocessJobAction>(JA) || Output.getType() != types::TY_PP_Asm) {
     Args.AddLastArg(CmdArgs, options::OPT_faltivec);
+    Args.AddLastArg(CmdArgs, options::OPT_fzvector);
+  }
   Args.AddLastArg(CmdArgs, options::OPT_fdiagnostics_show_template_tree);
   Args.AddLastArg(CmdArgs, options::OPT_fno_elide_type);
 
@@ -4096,6 +4098,12 @@
                                                        << "ppc/ppc64/ppc64le";
   }
 
+  // -fzvector is incompatible with -faltivec.
+  if (Arg *A = Args.getLastArg(options::OPT_fzvector))
+    if (Args.hasArg(options::OPT_faltivec))
+      D.Diag(diag::err_drv_argument_not_allowed_with) << A->getAsString(Args)
+                                                      << "-faltivec";
+
   if (getToolChain().SupportsProfiling())
     Args.AddLastArg(CmdArgs, options::OPT_pg);
 
diff --git a/clang/lib/Frontend/CompilerInvocation.cpp b/clang/lib/Frontend/CompilerInvocation.cpp
index f325144..fbeba09e 100644
--- a/clang/lib/Frontend/CompilerInvocation.cpp
+++ b/clang/lib/Frontend/CompilerInvocation.cpp
@@ -1260,6 +1260,7 @@
   // OpenCL has some additional defaults.
   if (Opts.OpenCL) {
     Opts.AltiVec = 0;
+    Opts.ZVector = 0;
     Opts.CXXOperatorNames = 1;
     Opts.LaxVectorConversions = 0;
     Opts.DefaultFPContract = 1;
@@ -1448,6 +1449,9 @@
   if (Args.hasArg(OPT_faltivec))
     Opts.AltiVec = 1;
 
+  if (Args.hasArg(OPT_fzvector))
+    Opts.ZVector = 1;
+
   if (Args.hasArg(OPT_pthread))
     Opts.POSIXThreads = 1;
 
diff --git a/clang/lib/Headers/CMakeLists.txt b/clang/lib/Headers/CMakeLists.txt
index 87afc60..7de5fbe 100644
--- a/clang/lib/Headers/CMakeLists.txt
+++ b/clang/lib/Headers/CMakeLists.txt
@@ -59,6 +59,7 @@
   unwind.h
   vadefs.h
   varargs.h
+  vecintrin.h
   __wmmintrin_aes.h
   wmmintrin.h
   __wmmintrin_pclmul.h
diff --git a/clang/lib/Headers/module.modulemap b/clang/lib/Headers/module.modulemap
index 0fc70a8..b147e89 100644
--- a/clang/lib/Headers/module.modulemap
+++ b/clang/lib/Headers/module.modulemap
@@ -158,6 +158,11 @@
       header "htmintrin.h"
       header "htmxlintrin.h"
     }
+
+    explicit module zvector {
+      requires zvector, vx
+      header "vecintrin.h"
+    }
   }
 }
 
diff --git a/clang/lib/Headers/s390intrin.h b/clang/lib/Headers/s390intrin.h
index b209895..d51274c 100644
--- a/clang/lib/Headers/s390intrin.h
+++ b/clang/lib/Headers/s390intrin.h
@@ -32,4 +32,8 @@
 #include <htmintrin.h>
 #endif
 
+#ifdef __VEC__
+#include <vecintrin.h>
+#endif
+
 #endif /* __S390INTRIN_H*/
diff --git a/clang/lib/Headers/vecintrin.h b/clang/lib/Headers/vecintrin.h
new file mode 100644
index 0000000..ca7acb4
--- /dev/null
+++ b/clang/lib/Headers/vecintrin.h
@@ -0,0 +1,8946 @@
+/*===---- vecintrin.h - Vector intrinsics ----------------------------------===
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ *
+ *===-----------------------------------------------------------------------===
+ */
+
+#if defined(__s390x__) && defined(__VEC__)
+
+#define __ATTRS_ai __attribute__((__always_inline__))
+#define __ATTRS_o __attribute__((__overloadable__))
+#define __ATTRS_o_ai __attribute__((__overloadable__, __always_inline__))
+
+#define __constant(PARM) \
+  __attribute__((__enable_if__ ((PARM) == (PARM), \
+     "argument must be a constant integer")))
+#define __constant_range(PARM, LOW, HIGH) \
+  __attribute__((__enable_if__ ((PARM) >= (LOW) && (PARM) <= (HIGH), \
+     "argument must be a constant integer from " #LOW " to " #HIGH)))
+#define __constant_pow2_range(PARM, LOW, HIGH) \
+  __attribute__((__enable_if__ ((PARM) >= (LOW) && (PARM) <= (HIGH) && \
+                                ((PARM) & ((PARM) - 1)) == 0, \
+     "argument must be a constant power of 2 from " #LOW " to " #HIGH)))
+
+/*-- __lcbb -----------------------------------------------------------------*/
+
+extern __ATTRS_o unsigned int
+__lcbb(const void *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+#define __lcbb(X, Y) ((__typeof__((__lcbb)((X), (Y)))) \
+  __builtin_s390_lcbb((X), __builtin_constant_p((Y))? \
+                           ((Y) == 64 ? 0 : \
+                            (Y) == 128 ? 1 : \
+                            (Y) == 256 ? 2 : \
+                            (Y) == 512 ? 3 : \
+                            (Y) == 1024 ? 4 : \
+                            (Y) == 2048 ? 5 : \
+                            (Y) == 4096 ? 6 : 0) : 0))
+
+/*-- vec_extract ------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai signed char
+vec_extract(vector signed char __vec, int __index) {
+  return __vec[__index & 15];
+}
+
+static inline __ATTRS_o_ai unsigned char
+vec_extract(vector bool char __vec, int __index) {
+  return __vec[__index & 15];
+}
+
+static inline __ATTRS_o_ai unsigned char
+vec_extract(vector unsigned char __vec, int __index) {
+  return __vec[__index & 15];
+}
+
+static inline __ATTRS_o_ai signed short
+vec_extract(vector signed short __vec, int __index) {
+  return __vec[__index & 7];
+}
+
+static inline __ATTRS_o_ai unsigned short
+vec_extract(vector bool short __vec, int __index) {
+  return __vec[__index & 7];
+}
+
+static inline __ATTRS_o_ai unsigned short
+vec_extract(vector unsigned short __vec, int __index) {
+  return __vec[__index & 7];
+}
+
+static inline __ATTRS_o_ai signed int
+vec_extract(vector signed int __vec, int __index) {
+  return __vec[__index & 3];
+}
+
+static inline __ATTRS_o_ai unsigned int
+vec_extract(vector bool int __vec, int __index) {
+  return __vec[__index & 3];
+}
+
+static inline __ATTRS_o_ai unsigned int
+vec_extract(vector unsigned int __vec, int __index) {
+  return __vec[__index & 3];
+}
+
+static inline __ATTRS_o_ai signed long long
+vec_extract(vector signed long long __vec, int __index) {
+  return __vec[__index & 1];
+}
+
+static inline __ATTRS_o_ai unsigned long long
+vec_extract(vector bool long long __vec, int __index) {
+  return __vec[__index & 1];
+}
+
+static inline __ATTRS_o_ai unsigned long long
+vec_extract(vector unsigned long long __vec, int __index) {
+  return __vec[__index & 1];
+}
+
+static inline __ATTRS_o_ai double
+vec_extract(vector double __vec, int __index) {
+  return __vec[__index & 1];
+}
+
+/*-- vec_insert -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_insert(signed char __scalar, vector signed char __vec, int __index) {
+  __vec[__index & 15] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_insert(unsigned char __scalar, vector bool char __vec, int __index) {
+  vector unsigned char __newvec = (vector unsigned char)__vec;
+  __newvec[__index & 15] = (unsigned char)__scalar;
+  return __newvec;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_insert(unsigned char __scalar, vector unsigned char __vec, int __index) {
+  __vec[__index & 15] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_insert(signed short __scalar, vector signed short __vec, int __index) {
+  __vec[__index & 7] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_insert(unsigned short __scalar, vector bool short __vec, int __index) {
+  vector unsigned short __newvec = (vector unsigned short)__vec;
+  __newvec[__index & 7] = (unsigned short)__scalar;
+  return __newvec;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_insert(unsigned short __scalar, vector unsigned short __vec, int __index) {
+  __vec[__index & 7] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_insert(signed int __scalar, vector signed int __vec, int __index) {
+  __vec[__index & 3] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_insert(unsigned int __scalar, vector bool int __vec, int __index) {
+  vector unsigned int __newvec = (vector unsigned int)__vec;
+  __newvec[__index & 3] = __scalar;
+  return __newvec;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_insert(unsigned int __scalar, vector unsigned int __vec, int __index) {
+  __vec[__index & 3] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_insert(signed long long __scalar, vector signed long long __vec,
+           int __index) {
+  __vec[__index & 1] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_insert(unsigned long long __scalar, vector bool long long __vec,
+           int __index) {
+  vector unsigned long long __newvec = (vector unsigned long long)__vec;
+  __newvec[__index & 1] = __scalar;
+  return __newvec;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_insert(unsigned long long __scalar, vector unsigned long long __vec,
+           int __index) {
+  __vec[__index & 1] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector double
+vec_insert(double __scalar, vector double __vec, int __index) {
+  __vec[__index & 1] = __scalar;
+  return __vec;
+}
+
+/*-- vec_promote ------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_promote(signed char __scalar, int __index) {
+  const vector signed char __zero = (vector signed char)0;
+  vector signed char __vec = __builtin_shufflevector(__zero, __zero,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1);
+  __vec[__index & 15] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_promote(unsigned char __scalar, int __index) {
+  const vector unsigned char __zero = (vector unsigned char)0;
+  vector unsigned char __vec = __builtin_shufflevector(__zero, __zero,
+    -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1);
+  __vec[__index & 15] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_promote(signed short __scalar, int __index) {
+  const vector signed short __zero = (vector signed short)0;
+  vector signed short __vec = __builtin_shufflevector(__zero, __zero,
+                                -1, -1, -1, -1, -1, -1, -1, -1);
+  __vec[__index & 7] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_promote(unsigned short __scalar, int __index) {
+  const vector unsigned short __zero = (vector unsigned short)0;
+  vector unsigned short __vec = __builtin_shufflevector(__zero, __zero,
+                                  -1, -1, -1, -1, -1, -1, -1, -1);
+  __vec[__index & 7] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_promote(signed int __scalar, int __index) {
+  const vector signed int __zero = (vector signed int)0;
+  vector signed int __vec = __builtin_shufflevector(__zero, __zero,
+                                                    -1, -1, -1, -1);
+  __vec[__index & 3] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_promote(unsigned int __scalar, int __index) {
+  const vector unsigned int __zero = (vector unsigned int)0;
+  vector unsigned int __vec = __builtin_shufflevector(__zero, __zero,
+                                                      -1, -1, -1, -1);
+  __vec[__index & 3] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_promote(signed long long __scalar, int __index) {
+  const vector signed long long __zero = (vector signed long long)0;
+  vector signed long long __vec = __builtin_shufflevector(__zero, __zero,
+                                                          -1, -1);
+  __vec[__index & 1] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_promote(unsigned long long __scalar, int __index) {
+  const vector unsigned long long __zero = (vector unsigned long long)0;
+  vector unsigned long long __vec = __builtin_shufflevector(__zero, __zero,
+                                                            -1, -1);
+  __vec[__index & 1] = __scalar;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector double
+vec_promote(double __scalar, int __index) {
+  const vector double __zero = (vector double)0;
+  vector double __vec = __builtin_shufflevector(__zero, __zero, -1, -1);
+  __vec[__index & 1] = __scalar;
+  return __vec;
+}
+
+/*-- vec_insert_and_zero ----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_insert_and_zero(const signed char *__ptr) {
+  vector signed char __vec = (vector signed char)0;
+  __vec[7] = *__ptr;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_insert_and_zero(const unsigned char *__ptr) {
+  vector unsigned char __vec = (vector unsigned char)0;
+  __vec[7] = *__ptr;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_insert_and_zero(const signed short *__ptr) {
+  vector signed short __vec = (vector signed short)0;
+  __vec[3] = *__ptr;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_insert_and_zero(const unsigned short *__ptr) {
+  vector unsigned short __vec = (vector unsigned short)0;
+  __vec[3] = *__ptr;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_insert_and_zero(const signed int *__ptr) {
+  vector signed int __vec = (vector signed int)0;
+  __vec[1] = *__ptr;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_insert_and_zero(const unsigned int *__ptr) {
+  vector unsigned int __vec = (vector unsigned int)0;
+  __vec[1] = *__ptr;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_insert_and_zero(const signed long long *__ptr) {
+  vector signed long long __vec = (vector signed long long)0;
+  __vec[0] = *__ptr;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_insert_and_zero(const unsigned long long *__ptr) {
+  vector unsigned long long __vec = (vector unsigned long long)0;
+  __vec[0] = *__ptr;
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector double
+vec_insert_and_zero(const double *__ptr) {
+  vector double __vec = (vector double)0;
+  __vec[0] = *__ptr;
+  return __vec;
+}
+
+/*-- vec_perm ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_perm(vector signed char __a, vector signed char __b,
+         vector unsigned char __c) {
+  return (vector signed char)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_perm(vector unsigned char __a, vector unsigned char __b,
+         vector unsigned char __c) {
+  return (vector unsigned char)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_perm(vector bool char __a, vector bool char __b,
+         vector unsigned char __c) {
+  return (vector bool char)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_perm(vector signed short __a, vector signed short __b,
+         vector unsigned char __c) {
+  return (vector signed short)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_perm(vector unsigned short __a, vector unsigned short __b,
+         vector unsigned char __c) {
+  return (vector unsigned short)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_perm(vector bool short __a, vector bool short __b,
+         vector unsigned char __c) {
+  return (vector bool short)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_perm(vector signed int __a, vector signed int __b,
+         vector unsigned char __c) {
+  return (vector signed int)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_perm(vector unsigned int __a, vector unsigned int __b,
+         vector unsigned char __c) {
+  return (vector unsigned int)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_perm(vector bool int __a, vector bool int __b,
+         vector unsigned char __c) {
+  return (vector bool int)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_perm(vector signed long long __a, vector signed long long __b,
+         vector unsigned char __c) {
+  return (vector signed long long)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_perm(vector unsigned long long __a, vector unsigned long long __b,
+         vector unsigned char __c) {
+  return (vector unsigned long long)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_perm(vector bool long long __a, vector bool long long __b,
+         vector unsigned char __c) {
+  return (vector bool long long)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_perm(vector double __a, vector double __b,
+         vector unsigned char __c) {
+  return (vector double)__builtin_s390_vperm(
+           (vector unsigned char)__a, (vector unsigned char)__b, __c);
+}
+
+/*-- vec_permi --------------------------------------------------------------*/
+
+extern __ATTRS_o vector signed long long
+vec_permi(vector signed long long __a, vector signed long long __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector unsigned long long
+vec_permi(vector unsigned long long __a, vector unsigned long long __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector bool long long
+vec_permi(vector bool long long __a, vector bool long long __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector double
+vec_permi(vector double __a, vector double __b, int __c)
+  __constant_range(__c, 0, 3);
+
+#define vec_permi(X, Y, Z) ((__typeof__((vec_permi)((X), (Y), (Z)))) \
+  __builtin_s390_vpdi((vector unsigned long long)(X), \
+                      (vector unsigned long long)(Y), \
+                      (((Z) & 2) << 1) | ((Z) & 1)))
+
+/*-- vec_sel ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_sel(vector signed char __a, vector signed char __b,
+        vector unsigned char __c) {
+  return ((vector signed char)__c & __b) | (~(vector signed char)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_sel(vector signed char __a, vector signed char __b, vector bool char __c) {
+  return ((vector signed char)__c & __b) | (~(vector signed char)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_sel(vector bool char __a, vector bool char __b, vector unsigned char __c) {
+  return ((vector bool char)__c & __b) | (~(vector bool char)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_sel(vector bool char __a, vector bool char __b, vector bool char __c) {
+  return (__c & __b) | (~__c & __a);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sel(vector unsigned char __a, vector unsigned char __b,
+        vector unsigned char __c) {
+  return (__c & __b) | (~__c & __a);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sel(vector unsigned char __a, vector unsigned char __b,
+        vector bool char __c) {
+  return ((vector unsigned char)__c & __b) | (~(vector unsigned char)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_sel(vector signed short __a, vector signed short __b,
+        vector unsigned short __c) {
+  return ((vector signed short)__c & __b) | (~(vector signed short)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_sel(vector signed short __a, vector signed short __b,
+        vector bool short __c) {
+  return ((vector signed short)__c & __b) | (~(vector signed short)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_sel(vector bool short __a, vector bool short __b,
+        vector unsigned short __c) {
+  return ((vector bool short)__c & __b) | (~(vector bool short)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_sel(vector bool short __a, vector bool short __b, vector bool short __c) {
+  return (__c & __b) | (~__c & __a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_sel(vector unsigned short __a, vector unsigned short __b,
+        vector unsigned short __c) {
+  return (__c & __b) | (~__c & __a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_sel(vector unsigned short __a, vector unsigned short __b,
+        vector bool short __c) {
+  return (((vector unsigned short)__c & __b) |
+          (~(vector unsigned short)__c & __a));
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_sel(vector signed int __a, vector signed int __b,
+        vector unsigned int __c) {
+  return ((vector signed int)__c & __b) | (~(vector signed int)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_sel(vector signed int __a, vector signed int __b, vector bool int __c) {
+  return ((vector signed int)__c & __b) | (~(vector signed int)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_sel(vector bool int __a, vector bool int __b, vector unsigned int __c) {
+  return ((vector bool int)__c & __b) | (~(vector bool int)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_sel(vector bool int __a, vector bool int __b, vector bool int __c) {
+  return (__c & __b) | (~__c & __a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sel(vector unsigned int __a, vector unsigned int __b,
+        vector unsigned int __c) {
+  return (__c & __b) | (~__c & __a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sel(vector unsigned int __a, vector unsigned int __b, vector bool int __c) {
+  return ((vector unsigned int)__c & __b) | (~(vector unsigned int)__c & __a);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_sel(vector signed long long __a, vector signed long long __b,
+        vector unsigned long long __c) {
+  return (((vector signed long long)__c & __b) |
+          (~(vector signed long long)__c & __a));
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_sel(vector signed long long __a, vector signed long long __b,
+        vector bool long long __c) {
+  return (((vector signed long long)__c & __b) |
+          (~(vector signed long long)__c & __a));
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_sel(vector bool long long __a, vector bool long long __b,
+        vector unsigned long long __c) {
+  return (((vector bool long long)__c & __b) |
+          (~(vector bool long long)__c & __a));
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_sel(vector bool long long __a, vector bool long long __b,
+        vector bool long long __c) {
+  return (__c & __b) | (~__c & __a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sel(vector unsigned long long __a, vector unsigned long long __b,
+        vector unsigned long long __c) {
+  return (__c & __b) | (~__c & __a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sel(vector unsigned long long __a, vector unsigned long long __b,
+        vector bool long long __c) {
+  return (((vector unsigned long long)__c & __b) |
+          (~(vector unsigned long long)__c & __a));
+}
+
+static inline __ATTRS_o_ai vector double
+vec_sel(vector double __a, vector double __b, vector unsigned long long __c) {
+  return (vector double)((__c & (vector unsigned long long)__b) |
+                         (~__c & (vector unsigned long long)__a));
+}
+
+static inline __ATTRS_o_ai vector double
+vec_sel(vector double __a, vector double __b, vector bool long long __c) {
+  vector unsigned long long __ac = (vector unsigned long long)__a;
+  vector unsigned long long __bc = (vector unsigned long long)__b;
+  vector unsigned long long __cc = (vector unsigned long long)__c;
+  return (vector double)((__cc & __bc) | (~__cc & __ac));
+}
+
+/*-- vec_gather_element -----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed int
+vec_gather_element(vector signed int __vec, vector unsigned int __offset,
+                   const signed int *__ptr, int __index)
+  __constant_range(__index, 0, 3) {
+  __vec[__index] = *(const signed int *)(
+    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_gather_element(vector bool int __vec, vector unsigned int __offset,
+                   const unsigned int *__ptr, int __index)
+  __constant_range(__index, 0, 3) {
+  __vec[__index] = *(const unsigned int *)(
+    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_gather_element(vector unsigned int __vec, vector unsigned int __offset,
+                   const unsigned int *__ptr, int __index)
+  __constant_range(__index, 0, 3) {
+  __vec[__index] = *(const unsigned int *)(
+    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_gather_element(vector signed long long __vec,
+                   vector unsigned long long __offset,
+                   const signed long long *__ptr, int __index)
+  __constant_range(__index, 0, 1) {
+  __vec[__index] = *(const signed long long *)(
+    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_gather_element(vector bool long long __vec,
+                   vector unsigned long long __offset,
+                   const unsigned long long *__ptr, int __index)
+  __constant_range(__index, 0, 1) {
+  __vec[__index] = *(const unsigned long long *)(
+    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_gather_element(vector unsigned long long __vec,
+                   vector unsigned long long __offset,
+                   const unsigned long long *__ptr, int __index)
+  __constant_range(__index, 0, 1) {
+  __vec[__index] = *(const unsigned long long *)(
+    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
+  return __vec;
+}
+
+static inline __ATTRS_o_ai vector double
+vec_gather_element(vector double __vec, vector unsigned long long __offset,
+                   const double *__ptr, int __index)
+  __constant_range(__index, 0, 1) {
+  __vec[__index] = *(const double *)(
+    (__INTPTR_TYPE__)__ptr + (__INTPTR_TYPE__)__offset[__index]);
+  return __vec;
+}
+
+/*-- vec_scatter_element ----------------------------------------------------*/
+
+static inline __ATTRS_o_ai void
+vec_scatter_element(vector signed int __vec, vector unsigned int __offset,
+                    signed int *__ptr, int __index)
+  __constant_range(__index, 0, 3) {
+  *(signed int *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
+    __vec[__index];
+}
+
+static inline __ATTRS_o_ai void
+vec_scatter_element(vector bool int __vec, vector unsigned int __offset,
+                    unsigned int *__ptr, int __index)
+  __constant_range(__index, 0, 3) {
+  *(unsigned int *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
+    __vec[__index];
+}
+
+static inline __ATTRS_o_ai void
+vec_scatter_element(vector unsigned int __vec, vector unsigned int __offset,
+                    unsigned int *__ptr, int __index)
+  __constant_range(__index, 0, 3) {
+  *(unsigned int *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
+    __vec[__index];
+}
+
+static inline __ATTRS_o_ai void
+vec_scatter_element(vector signed long long __vec,
+                    vector unsigned long long __offset,
+                    signed long long *__ptr, int __index)
+  __constant_range(__index, 0, 1) {
+  *(signed long long *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
+    __vec[__index];
+}
+
+static inline __ATTRS_o_ai void
+vec_scatter_element(vector bool long long __vec,
+                    vector unsigned long long __offset,
+                    unsigned long long *__ptr, int __index)
+  __constant_range(__index, 0, 1) {
+  *(unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
+    __vec[__index];
+}
+
+static inline __ATTRS_o_ai void
+vec_scatter_element(vector unsigned long long __vec,
+                    vector unsigned long long __offset,
+                    unsigned long long *__ptr, int __index)
+  __constant_range(__index, 0, 1) {
+  *(unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
+    __vec[__index];
+}
+
+static inline __ATTRS_o_ai void
+vec_scatter_element(vector double __vec, vector unsigned long long __offset,
+                    double *__ptr, int __index)
+  __constant_range(__index, 0, 1) {
+  *(double *)((__INTPTR_TYPE__)__ptr + __offset[__index]) =
+    __vec[__index];
+}
+
+/*-- vec_xld2 ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_xld2(long __offset, const signed char *__ptr) {
+  return *(const vector signed char *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_xld2(long __offset, const unsigned char *__ptr) {
+  return *(const vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_xld2(long __offset, const signed short *__ptr) {
+  return *(const vector signed short *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_xld2(long __offset, const unsigned short *__ptr) {
+  return *(const vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_xld2(long __offset, const signed int *__ptr) {
+  return *(const vector signed int *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_xld2(long __offset, const unsigned int *__ptr) {
+  return *(const vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_xld2(long __offset, const signed long long *__ptr) {
+  return *(const vector signed long long *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_xld2(long __offset, const unsigned long long *__ptr) {
+  return *(const vector unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_xld2(long __offset, const double *__ptr) {
+  return *(const vector double *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+/*-- vec_xlw4 ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_xlw4(long __offset, const signed char *__ptr) {
+  return *(const vector signed char *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_xlw4(long __offset, const unsigned char *__ptr) {
+  return *(const vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_xlw4(long __offset, const signed short *__ptr) {
+  return *(const vector signed short *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_xlw4(long __offset, const unsigned short *__ptr) {
+  return *(const vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_xlw4(long __offset, const signed int *__ptr) {
+  return *(const vector signed int *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_xlw4(long __offset, const unsigned int *__ptr) {
+  return *(const vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset);
+}
+
+/*-- vec_xstd2 --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector signed char __vec, long __offset, signed char *__ptr) {
+  *(vector signed char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector unsigned char __vec, long __offset, unsigned char *__ptr) {
+  *(vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector signed short __vec, long __offset, signed short *__ptr) {
+  *(vector signed short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector unsigned short __vec, long __offset, unsigned short *__ptr) {
+  *(vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector signed int __vec, long __offset, signed int *__ptr) {
+  *(vector signed int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector unsigned int __vec, long __offset, unsigned int *__ptr) {
+  *(vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector signed long long __vec, long __offset,
+          signed long long *__ptr) {
+  *(vector signed long long *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector unsigned long long __vec, long __offset,
+          unsigned long long *__ptr) {
+  *(vector unsigned long long *)((__INTPTR_TYPE__)__ptr + __offset) =
+    __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstd2(vector double __vec, long __offset, double *__ptr) {
+  *(vector double *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+/*-- vec_xstw4 --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai void
+vec_xstw4(vector signed char __vec, long __offset, signed char *__ptr) {
+  *(vector signed char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstw4(vector unsigned char __vec, long __offset, unsigned char *__ptr) {
+  *(vector unsigned char *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstw4(vector signed short __vec, long __offset, signed short *__ptr) {
+  *(vector signed short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstw4(vector unsigned short __vec, long __offset, unsigned short *__ptr) {
+  *(vector unsigned short *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstw4(vector signed int __vec, long __offset, signed int *__ptr) {
+  *(vector signed int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+static inline __ATTRS_o_ai void
+vec_xstw4(vector unsigned int __vec, long __offset, unsigned int *__ptr) {
+  *(vector unsigned int *)((__INTPTR_TYPE__)__ptr + __offset) = __vec;
+}
+
+/*-- vec_load_bndry ---------------------------------------------------------*/
+
+extern __ATTRS_o vector signed char
+vec_load_bndry(const signed char *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+extern __ATTRS_o vector unsigned char
+vec_load_bndry(const unsigned char *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+extern __ATTRS_o vector signed short
+vec_load_bndry(const signed short *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+extern __ATTRS_o vector unsigned short
+vec_load_bndry(const unsigned short *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+extern __ATTRS_o vector signed int
+vec_load_bndry(const signed int *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+extern __ATTRS_o vector unsigned int
+vec_load_bndry(const unsigned int *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+extern __ATTRS_o vector signed long long
+vec_load_bndry(const signed long long *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+extern __ATTRS_o vector unsigned long long
+vec_load_bndry(const unsigned long long *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+extern __ATTRS_o vector double
+vec_load_bndry(const double *__ptr, unsigned short __len)
+  __constant_pow2_range(__len, 64, 4096);
+
+#define vec_load_bndry(X, Y) ((__typeof__((vec_load_bndry)((X), (Y)))) \
+  __builtin_s390_vlbb((X), ((Y) == 64 ? 0 : \
+                            (Y) == 128 ? 1 : \
+                            (Y) == 256 ? 2 : \
+                            (Y) == 512 ? 3 : \
+                            (Y) == 1024 ? 4 : \
+                            (Y) == 2048 ? 5 : \
+                            (Y) == 4096 ? 6 : -1)))
+
+/*-- vec_load_len -----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_load_len(const signed char *__ptr, unsigned int __len) {
+  return (vector signed char)__builtin_s390_vll(__len, __ptr);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_load_len(const unsigned char *__ptr, unsigned int __len) {
+  return (vector unsigned char)__builtin_s390_vll(__len, __ptr);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_load_len(const signed short *__ptr, unsigned int __len) {
+  return (vector signed short)__builtin_s390_vll(__len, __ptr);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_load_len(const unsigned short *__ptr, unsigned int __len) {
+  return (vector unsigned short)__builtin_s390_vll(__len, __ptr);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_load_len(const signed int *__ptr, unsigned int __len) {
+  return (vector signed int)__builtin_s390_vll(__len, __ptr);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_load_len(const unsigned int *__ptr, unsigned int __len) {
+  return (vector unsigned int)__builtin_s390_vll(__len, __ptr);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_load_len(const signed long long *__ptr, unsigned int __len) {
+  return (vector signed long long)__builtin_s390_vll(__len, __ptr);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_load_len(const unsigned long long *__ptr, unsigned int __len) {
+  return (vector unsigned long long)__builtin_s390_vll(__len, __ptr);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_load_len(const double *__ptr, unsigned int __len) {
+  return (vector double)__builtin_s390_vll(__len, __ptr);
+}
+
+/*-- vec_store_len ----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector signed char __vec, signed char *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector unsigned char __vec, unsigned char *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector signed short __vec, signed short *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector unsigned short __vec, unsigned short *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector signed int __vec, signed int *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector unsigned int __vec, unsigned int *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector signed long long __vec, signed long long *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector unsigned long long __vec, unsigned long long *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+static inline __ATTRS_o_ai void
+vec_store_len(vector double __vec, double *__ptr,
+              unsigned int __len) {
+  __builtin_s390_vstl((vector signed char)__vec, __len, __ptr);
+}
+
+/*-- vec_load_pair ----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed long long
+vec_load_pair(signed long long __a, signed long long __b) {
+  return (vector signed long long)(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_load_pair(unsigned long long __a, unsigned long long __b) {
+  return (vector unsigned long long)(__a, __b);
+}
+
+/*-- vec_genmask ------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_genmask(unsigned short __mask)
+  __constant(__mask) {
+  return (vector unsigned char)(
+    __mask & 0x8000 ? 0xff : 0,
+    __mask & 0x4000 ? 0xff : 0,
+    __mask & 0x2000 ? 0xff : 0,
+    __mask & 0x1000 ? 0xff : 0,
+    __mask & 0x0800 ? 0xff : 0,
+    __mask & 0x0400 ? 0xff : 0,
+    __mask & 0x0200 ? 0xff : 0,
+    __mask & 0x0100 ? 0xff : 0,
+    __mask & 0x0080 ? 0xff : 0,
+    __mask & 0x0040 ? 0xff : 0,
+    __mask & 0x0020 ? 0xff : 0,
+    __mask & 0x0010 ? 0xff : 0,
+    __mask & 0x0008 ? 0xff : 0,
+    __mask & 0x0004 ? 0xff : 0,
+    __mask & 0x0002 ? 0xff : 0,
+    __mask & 0x0001 ? 0xff : 0);
+}
+
+/*-- vec_genmasks_* ---------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_genmasks_8(unsigned char __first, unsigned char __last)
+  __constant(__first) __constant(__last) {
+  unsigned char __bit1 = __first & 7;
+  unsigned char __bit2 = __last & 7;
+  unsigned char __mask1 = (unsigned char)(1U << (7 - __bit1) << 1) - 1;
+  unsigned char __mask2 = (unsigned char)(1U << (7 - __bit2)) - 1;
+  unsigned char __value = (__bit1 <= __bit2 ?
+                           __mask1 & ~__mask2 :
+                           __mask1 | ~__mask2);
+  return (vector unsigned char)__value;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_genmasks_16(unsigned char __first, unsigned char __last)
+  __constant(__first) __constant(__last) {
+  unsigned char __bit1 = __first & 15;
+  unsigned char __bit2 = __last & 15;
+  unsigned short __mask1 = (unsigned short)(1U << (15 - __bit1) << 1) - 1;
+  unsigned short __mask2 = (unsigned short)(1U << (15 - __bit2)) - 1;
+  unsigned short __value = (__bit1 <= __bit2 ?
+                            __mask1 & ~__mask2 :
+                            __mask1 | ~__mask2);
+  return (vector unsigned short)__value;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_genmasks_32(unsigned char __first, unsigned char __last)
+  __constant(__first) __constant(__last) {
+  unsigned char __bit1 = __first & 31;
+  unsigned char __bit2 = __last & 31;
+  unsigned int __mask1 = (1U << (31 - __bit1) << 1) - 1;
+  unsigned int __mask2 = (1U << (31 - __bit2)) - 1;
+  unsigned int __value = (__bit1 <= __bit2 ?
+                          __mask1 & ~__mask2 :
+                          __mask1 | ~__mask2);
+  return (vector unsigned int)__value;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_genmasks_64(unsigned char __first, unsigned char __last)
+  __constant(__first) __constant(__last) {
+  unsigned char __bit1 = __first & 63;
+  unsigned char __bit2 = __last & 63;
+  unsigned long long __mask1 = (1ULL << (63 - __bit1) << 1) - 1;
+  unsigned long long __mask2 = (1ULL << (63 - __bit2)) - 1;
+  unsigned long long __value = (__bit1 <= __bit2 ?
+                                __mask1 & ~__mask2 :
+                                __mask1 | ~__mask2);
+  return (vector unsigned long long)__value;
+}
+
+/*-- vec_splat --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_splat(vector signed char __vec, int __index)
+  __constant_range(__index, 0, 15) {
+  return (vector signed char)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_splat(vector bool char __vec, int __index)
+  __constant_range(__index, 0, 15) {
+  return (vector bool char)(vector unsigned char)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_splat(vector unsigned char __vec, int __index)
+  __constant_range(__index, 0, 15) {
+  return (vector unsigned char)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_splat(vector signed short __vec, int __index)
+  __constant_range(__index, 0, 7) {
+  return (vector signed short)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_splat(vector bool short __vec, int __index)
+  __constant_range(__index, 0, 7) {
+  return (vector bool short)(vector unsigned short)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_splat(vector unsigned short __vec, int __index)
+  __constant_range(__index, 0, 7) {
+  return (vector unsigned short)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_splat(vector signed int __vec, int __index)
+  __constant_range(__index, 0, 3) {
+  return (vector signed int)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_splat(vector bool int __vec, int __index)
+  __constant_range(__index, 0, 3) {
+  return (vector bool int)(vector unsigned int)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_splat(vector unsigned int __vec, int __index)
+  __constant_range(__index, 0, 3) {
+  return (vector unsigned int)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_splat(vector signed long long __vec, int __index)
+  __constant_range(__index, 0, 1) {
+  return (vector signed long long)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_splat(vector bool long long __vec, int __index)
+  __constant_range(__index, 0, 1) {
+  return (vector bool long long)(vector unsigned long long)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_splat(vector unsigned long long __vec, int __index)
+  __constant_range(__index, 0, 1) {
+  return (vector unsigned long long)__vec[__index];
+}
+
+static inline __ATTRS_o_ai vector double
+vec_splat(vector double __vec, int __index)
+  __constant_range(__index, 0, 1) {
+  return (vector double)__vec[__index];
+}
+
+/*-- vec_splat_s* -----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector signed char
+vec_splat_s8(signed char __scalar)
+  __constant(__scalar) {
+  return (vector signed char)__scalar;
+}
+
+static inline __ATTRS_ai vector signed short
+vec_splat_s16(signed short __scalar)
+  __constant(__scalar) {
+  return (vector signed short)__scalar;
+}
+
+static inline __ATTRS_ai vector signed int
+vec_splat_s32(signed short __scalar)
+  __constant(__scalar) {
+  return (vector signed int)(signed int)__scalar;
+}
+
+static inline __ATTRS_ai vector signed long long
+vec_splat_s64(signed short __scalar)
+  __constant(__scalar) {
+  return (vector signed long long)(signed long)__scalar;
+}
+
+/*-- vec_splat_u* -----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_splat_u8(unsigned char __scalar)
+  __constant(__scalar) {
+  return (vector unsigned char)__scalar;
+}
+
+static inline __ATTRS_ai vector unsigned short
+vec_splat_u16(unsigned short __scalar)
+  __constant(__scalar) {
+  return (vector unsigned short)__scalar;
+}
+
+static inline __ATTRS_ai vector unsigned int
+vec_splat_u32(signed short __scalar)
+  __constant(__scalar) {
+  return (vector unsigned int)(signed int)__scalar;
+}
+
+static inline __ATTRS_ai vector unsigned long long
+vec_splat_u64(signed short __scalar)
+  __constant(__scalar) {
+  return (vector unsigned long long)(signed long long)__scalar;
+}
+
+/*-- vec_splats -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_splats(signed char __scalar) {
+  return (vector signed char)__scalar;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_splats(unsigned char __scalar) {
+  return (vector unsigned char)__scalar;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_splats(signed short __scalar) {
+  return (vector signed short)__scalar;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_splats(unsigned short __scalar) {
+  return (vector unsigned short)__scalar;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_splats(signed int __scalar) {
+  return (vector signed int)__scalar;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_splats(unsigned int __scalar) {
+  return (vector unsigned int)__scalar;
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_splats(signed long long __scalar) {
+  return (vector signed long long)__scalar;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_splats(unsigned long long __scalar) {
+  return (vector unsigned long long)__scalar;
+}
+
+static inline __ATTRS_o_ai vector double
+vec_splats(double __scalar) {
+  return (vector double)__scalar;
+}
+
+/*-- vec_extend_s64 ---------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed long long
+vec_extend_s64(vector signed char __a) {
+  return (vector signed long long)(__a[7], __a[15]);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_extend_s64(vector signed short __a) {
+  return (vector signed long long)(__a[3], __a[7]);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_extend_s64(vector signed int __a) {
+  return (vector signed long long)(__a[1], __a[3]);
+}
+
+/*-- vec_mergeh -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_mergeh(vector signed char __a, vector signed char __b) {
+  return (vector signed char)(
+    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3],
+    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_mergeh(vector bool char __a, vector bool char __b) {
+  return (vector bool char)(
+    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3],
+    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_mergeh(vector unsigned char __a, vector unsigned char __b) {
+  return (vector unsigned char)(
+    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3],
+    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_mergeh(vector signed short __a, vector signed short __b) {
+  return (vector signed short)(
+    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3]);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_mergeh(vector bool short __a, vector bool short __b) {
+  return (vector bool short)(
+    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3]);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_mergeh(vector unsigned short __a, vector unsigned short __b) {
+  return (vector unsigned short)(
+    __a[0], __b[0], __a[1], __b[1], __a[2], __b[2], __a[3], __b[3]);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mergeh(vector signed int __a, vector signed int __b) {
+  return (vector signed int)(__a[0], __b[0], __a[1], __b[1]);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_mergeh(vector bool int __a, vector bool int __b) {
+  return (vector bool int)(__a[0], __b[0], __a[1], __b[1]);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_mergeh(vector unsigned int __a, vector unsigned int __b) {
+  return (vector unsigned int)(__a[0], __b[0], __a[1], __b[1]);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_mergeh(vector signed long long __a, vector signed long long __b) {
+  return (vector signed long long)(__a[0], __b[0]);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_mergeh(vector bool long long __a, vector bool long long __b) {
+  return (vector bool long long)(__a[0], __b[0]);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_mergeh(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector unsigned long long)(__a[0], __b[0]);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_mergeh(vector double __a, vector double __b) {
+  return (vector double)(__a[0], __b[0]);
+}
+
+/*-- vec_mergel -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_mergel(vector signed char __a, vector signed char __b) {
+  return (vector signed char)(
+    __a[8], __b[8], __a[9], __b[9], __a[10], __b[10], __a[11], __b[11],
+    __a[12], __b[12], __a[13], __b[13], __a[14], __b[14], __a[15], __b[15]);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_mergel(vector bool char __a, vector bool char __b) {
+  return (vector bool char)(
+    __a[8], __b[8], __a[9], __b[9], __a[10], __b[10], __a[11], __b[11],
+    __a[12], __b[12], __a[13], __b[13], __a[14], __b[14], __a[15], __b[15]);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_mergel(vector unsigned char __a, vector unsigned char __b) {
+  return (vector unsigned char)(
+    __a[8], __b[8], __a[9], __b[9], __a[10], __b[10], __a[11], __b[11],
+    __a[12], __b[12], __a[13], __b[13], __a[14], __b[14], __a[15], __b[15]);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_mergel(vector signed short __a, vector signed short __b) {
+  return (vector signed short)(
+    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_mergel(vector bool short __a, vector bool short __b) {
+  return (vector bool short)(
+    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_mergel(vector unsigned short __a, vector unsigned short __b) {
+  return (vector unsigned short)(
+    __a[4], __b[4], __a[5], __b[5], __a[6], __b[6], __a[7], __b[7]);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mergel(vector signed int __a, vector signed int __b) {
+  return (vector signed int)(__a[2], __b[2], __a[3], __b[3]);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_mergel(vector bool int __a, vector bool int __b) {
+  return (vector bool int)(__a[2], __b[2], __a[3], __b[3]);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_mergel(vector unsigned int __a, vector unsigned int __b) {
+  return (vector unsigned int)(__a[2], __b[2], __a[3], __b[3]);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_mergel(vector signed long long __a, vector signed long long __b) {
+  return (vector signed long long)(__a[1], __b[1]);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_mergel(vector bool long long __a, vector bool long long __b) {
+  return (vector bool long long)(__a[1], __b[1]);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_mergel(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector unsigned long long)(__a[1], __b[1]);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_mergel(vector double __a, vector double __b) {
+  return (vector double)(__a[1], __b[1]);
+}
+
+/*-- vec_pack ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_pack(vector signed short __a, vector signed short __b) {
+  vector signed char __ac = (vector signed char)__a;
+  vector signed char __bc = (vector signed char)__b;
+  return (vector signed char)(
+    __ac[1], __ac[3], __ac[5], __ac[7], __ac[9], __ac[11], __ac[13], __ac[15],
+    __bc[1], __bc[3], __bc[5], __bc[7], __bc[9], __bc[11], __bc[13], __bc[15]);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_pack(vector bool short __a, vector bool short __b) {
+  vector bool char __ac = (vector bool char)__a;
+  vector bool char __bc = (vector bool char)__b;
+  return (vector bool char)(
+    __ac[1], __ac[3], __ac[5], __ac[7], __ac[9], __ac[11], __ac[13], __ac[15],
+    __bc[1], __bc[3], __bc[5], __bc[7], __bc[9], __bc[11], __bc[13], __bc[15]);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_pack(vector unsigned short __a, vector unsigned short __b) {
+  vector unsigned char __ac = (vector unsigned char)__a;
+  vector unsigned char __bc = (vector unsigned char)__b;
+  return (vector unsigned char)(
+    __ac[1], __ac[3], __ac[5], __ac[7], __ac[9], __ac[11], __ac[13], __ac[15],
+    __bc[1], __bc[3], __bc[5], __bc[7], __bc[9], __bc[11], __bc[13], __bc[15]);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_pack(vector signed int __a, vector signed int __b) {
+  vector signed short __ac = (vector signed short)__a;
+  vector signed short __bc = (vector signed short)__b;
+  return (vector signed short)(
+    __ac[1], __ac[3], __ac[5], __ac[7],
+    __bc[1], __bc[3], __bc[5], __bc[7]);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_pack(vector bool int __a, vector bool int __b) {
+  vector bool short __ac = (vector bool short)__a;
+  vector bool short __bc = (vector bool short)__b;
+  return (vector bool short)(
+    __ac[1], __ac[3], __ac[5], __ac[7],
+    __bc[1], __bc[3], __bc[5], __bc[7]);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_pack(vector unsigned int __a, vector unsigned int __b) {
+  vector unsigned short __ac = (vector unsigned short)__a;
+  vector unsigned short __bc = (vector unsigned short)__b;
+  return (vector unsigned short)(
+    __ac[1], __ac[3], __ac[5], __ac[7],
+    __bc[1], __bc[3], __bc[5], __bc[7]);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_pack(vector signed long long __a, vector signed long long __b) {
+  vector signed int __ac = (vector signed int)__a;
+  vector signed int __bc = (vector signed int)__b;
+  return (vector signed int)(__ac[1], __ac[3], __bc[1], __bc[3]);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_pack(vector bool long long __a, vector bool long long __b) {
+  vector bool int __ac = (vector bool int)__a;
+  vector bool int __bc = (vector bool int)__b;
+  return (vector bool int)(__ac[1], __ac[3], __bc[1], __bc[3]);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_pack(vector unsigned long long __a, vector unsigned long long __b) {
+  vector unsigned int __ac = (vector unsigned int)__a;
+  vector unsigned int __bc = (vector unsigned int)__b;
+  return (vector unsigned int)(__ac[1], __ac[3], __bc[1], __bc[3]);
+}
+
+/*-- vec_packs --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_packs(vector signed short __a, vector signed short __b) {
+  return __builtin_s390_vpksh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_packs(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vpklsh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_packs(vector signed int __a, vector signed int __b) {
+  return __builtin_s390_vpksf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_packs(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vpklsf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_packs(vector signed long long __a, vector signed long long __b) {
+  return __builtin_s390_vpksg(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_packs(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vpklsg(__a, __b);
+}
+
+/*-- vec_packs_cc -----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_packs_cc(vector signed short __a, vector signed short __b, int *__cc) {
+  return __builtin_s390_vpkshs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_packs_cc(vector unsigned short __a, vector unsigned short __b, int *__cc) {
+  return __builtin_s390_vpklshs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_packs_cc(vector signed int __a, vector signed int __b, int *__cc) {
+  return __builtin_s390_vpksfs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_packs_cc(vector unsigned int __a, vector unsigned int __b, int *__cc) {
+  return __builtin_s390_vpklsfs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_packs_cc(vector signed long long __a, vector signed long long __b,
+             int *__cc) {
+  return __builtin_s390_vpksgs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_packs_cc(vector unsigned long long __a, vector unsigned long long __b,
+             int *__cc) {
+  return __builtin_s390_vpklsgs(__a, __b, __cc);
+}
+
+/*-- vec_packsu -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_packsu(vector signed short __a, vector signed short __b) {
+  const vector signed short __zero = (vector signed short)0;
+  return __builtin_s390_vpklsh(
+    (vector unsigned short)(__a >= __zero) & (vector unsigned short)__a,
+    (vector unsigned short)(__b >= __zero) & (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_packsu(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vpklsh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_packsu(vector signed int __a, vector signed int __b) {
+  const vector signed int __zero = (vector signed int)0;
+  return __builtin_s390_vpklsf(
+    (vector unsigned int)(__a >= __zero) & (vector unsigned int)__a,
+    (vector unsigned int)(__b >= __zero) & (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_packsu(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vpklsf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_packsu(vector signed long long __a, vector signed long long __b) {
+  const vector signed long long __zero = (vector signed long long)0;
+  return __builtin_s390_vpklsg(
+    (vector unsigned long long)(__a >= __zero) &
+    (vector unsigned long long)__a,
+    (vector unsigned long long)(__b >= __zero) &
+    (vector unsigned long long)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_packsu(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vpklsg(__a, __b);
+}
+
+/*-- vec_packsu_cc ----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_packsu_cc(vector unsigned short __a, vector unsigned short __b, int *__cc) {
+  return __builtin_s390_vpklshs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_packsu_cc(vector unsigned int __a, vector unsigned int __b, int *__cc) {
+  return __builtin_s390_vpklsfs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_packsu_cc(vector unsigned long long __a, vector unsigned long long __b,
+              int *__cc) {
+  return __builtin_s390_vpklsgs(__a, __b, __cc);
+}
+
+/*-- vec_unpackh ------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed short
+vec_unpackh(vector signed char __a) {
+  return __builtin_s390_vuphb(__a);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_unpackh(vector bool char __a) {
+  return (vector bool short)__builtin_s390_vuphb((vector signed char)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_unpackh(vector unsigned char __a) {
+  return __builtin_s390_vuplhb(__a);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_unpackh(vector signed short __a) {
+  return __builtin_s390_vuphh(__a);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_unpackh(vector bool short __a) {
+  return (vector bool int)__builtin_s390_vuphh((vector signed short)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_unpackh(vector unsigned short __a) {
+  return __builtin_s390_vuplhh(__a);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_unpackh(vector signed int __a) {
+  return __builtin_s390_vuphf(__a);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_unpackh(vector bool int __a) {
+  return (vector bool long long)__builtin_s390_vuphf((vector signed int)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_unpackh(vector unsigned int __a) {
+  return __builtin_s390_vuplhf(__a);
+}
+
+/*-- vec_unpackl ------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed short
+vec_unpackl(vector signed char __a) {
+  return __builtin_s390_vuplb(__a);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_unpackl(vector bool char __a) {
+  return (vector bool short)__builtin_s390_vuplb((vector signed char)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_unpackl(vector unsigned char __a) {
+  return __builtin_s390_vupllb(__a);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_unpackl(vector signed short __a) {
+  return __builtin_s390_vuplhw(__a);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_unpackl(vector bool short __a) {
+  return (vector bool int)__builtin_s390_vuplhw((vector signed short)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_unpackl(vector unsigned short __a) {
+  return __builtin_s390_vupllh(__a);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_unpackl(vector signed int __a) {
+  return __builtin_s390_vuplf(__a);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_unpackl(vector bool int __a) {
+  return (vector bool long long)__builtin_s390_vuplf((vector signed int)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_unpackl(vector unsigned int __a) {
+  return __builtin_s390_vupllf(__a);
+}
+
+/*-- vec_cmpeq --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpeq(vector bool char __a, vector bool char __b) {
+  return (vector bool char)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpeq(vector signed char __a, vector signed char __b) {
+  return (vector bool char)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpeq(vector unsigned char __a, vector unsigned char __b) {
+  return (vector bool char)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpeq(vector bool short __a, vector bool short __b) {
+  return (vector bool short)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpeq(vector signed short __a, vector signed short __b) {
+  return (vector bool short)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpeq(vector unsigned short __a, vector unsigned short __b) {
+  return (vector bool short)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpeq(vector bool int __a, vector bool int __b) {
+  return (vector bool int)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpeq(vector signed int __a, vector signed int __b) {
+  return (vector bool int)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpeq(vector unsigned int __a, vector unsigned int __b) {
+  return (vector bool int)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpeq(vector bool long long __a, vector bool long long __b) {
+  return (vector bool long long)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpeq(vector signed long long __a, vector signed long long __b) {
+  return (vector bool long long)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpeq(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector bool long long)(__a == __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpeq(vector double __a, vector double __b) {
+  return (vector bool long long)(__a == __b);
+}
+
+/*-- vec_cmpge --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpge(vector signed char __a, vector signed char __b) {
+  return (vector bool char)(__a >= __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpge(vector unsigned char __a, vector unsigned char __b) {
+  return (vector bool char)(__a >= __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpge(vector signed short __a, vector signed short __b) {
+  return (vector bool short)(__a >= __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpge(vector unsigned short __a, vector unsigned short __b) {
+  return (vector bool short)(__a >= __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpge(vector signed int __a, vector signed int __b) {
+  return (vector bool int)(__a >= __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpge(vector unsigned int __a, vector unsigned int __b) {
+  return (vector bool int)(__a >= __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpge(vector signed long long __a, vector signed long long __b) {
+  return (vector bool long long)(__a >= __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpge(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector bool long long)(__a >= __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpge(vector double __a, vector double __b) {
+  return (vector bool long long)(__a >= __b);
+}
+
+/*-- vec_cmpgt --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpgt(vector signed char __a, vector signed char __b) {
+  return (vector bool char)(__a > __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpgt(vector unsigned char __a, vector unsigned char __b) {
+  return (vector bool char)(__a > __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpgt(vector signed short __a, vector signed short __b) {
+  return (vector bool short)(__a > __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpgt(vector unsigned short __a, vector unsigned short __b) {
+  return (vector bool short)(__a > __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpgt(vector signed int __a, vector signed int __b) {
+  return (vector bool int)(__a > __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpgt(vector unsigned int __a, vector unsigned int __b) {
+  return (vector bool int)(__a > __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpgt(vector signed long long __a, vector signed long long __b) {
+  return (vector bool long long)(__a > __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpgt(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector bool long long)(__a > __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmpgt(vector double __a, vector double __b) {
+  return (vector bool long long)(__a > __b);
+}
+
+/*-- vec_cmple --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmple(vector signed char __a, vector signed char __b) {
+  return (vector bool char)(__a <= __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmple(vector unsigned char __a, vector unsigned char __b) {
+  return (vector bool char)(__a <= __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmple(vector signed short __a, vector signed short __b) {
+  return (vector bool short)(__a <= __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmple(vector unsigned short __a, vector unsigned short __b) {
+  return (vector bool short)(__a <= __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmple(vector signed int __a, vector signed int __b) {
+  return (vector bool int)(__a <= __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmple(vector unsigned int __a, vector unsigned int __b) {
+  return (vector bool int)(__a <= __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmple(vector signed long long __a, vector signed long long __b) {
+  return (vector bool long long)(__a <= __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmple(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector bool long long)(__a <= __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmple(vector double __a, vector double __b) {
+  return (vector bool long long)(__a <= __b);
+}
+
+/*-- vec_cmplt --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmplt(vector signed char __a, vector signed char __b) {
+  return (vector bool char)(__a < __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmplt(vector unsigned char __a, vector unsigned char __b) {
+  return (vector bool char)(__a < __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmplt(vector signed short __a, vector signed short __b) {
+  return (vector bool short)(__a < __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmplt(vector unsigned short __a, vector unsigned short __b) {
+  return (vector bool short)(__a < __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmplt(vector signed int __a, vector signed int __b) {
+  return (vector bool int)(__a < __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmplt(vector unsigned int __a, vector unsigned int __b) {
+  return (vector bool int)(__a < __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmplt(vector signed long long __a, vector signed long long __b) {
+  return (vector bool long long)(__a < __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmplt(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector bool long long)(__a < __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_cmplt(vector double __a, vector double __b) {
+  return (vector bool long long)(__a < __b);
+}
+
+/*-- vec_all_eq -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vceqbs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs(__a, (vector signed char)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vceqhs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs(__a, (vector signed short)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vceqfs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs(__a, (vector signed int)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs(__a, (vector signed long long)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_eq(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfcedbs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+/*-- vec_all_ne -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vceqbs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs(__a, (vector signed char)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vceqhs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs(__a, (vector signed short)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vceqfs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs(__a, (vector signed int)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs(__a, (vector signed long long)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ne(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfcedbs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+/*-- vec_all_ge -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchbs((vector signed char)__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__b, (vector signed char)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__b, (vector unsigned char)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__b,
+                        (vector unsigned char)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchhs((vector signed short)__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__b, (vector signed short)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__b, (vector unsigned short)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__b,
+                        (vector unsigned short)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchfs((vector signed int)__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__b, (vector signed int)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__b, (vector unsigned int)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__b,
+                        (vector unsigned int)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchgs((vector signed long long)__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__b, (vector signed long long)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__b, __a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__b, (vector unsigned long long)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__b,
+                        (vector unsigned long long)__a, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_ge(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchedbs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+/*-- vec_all_gt -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__a, (vector signed char)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs((vector signed char)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__a, (vector unsigned char)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__a,
+                        (vector unsigned char)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__a, (vector signed short)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs((vector signed short)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__a, (vector unsigned short)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__a,
+                        (vector unsigned short)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__a, (vector signed int)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs((vector signed int)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__a, (vector unsigned int)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__a,
+                        (vector unsigned int)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__a, (vector signed long long)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs((vector signed long long)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__a, (vector unsigned long long)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__a, __b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__a,
+                        (vector unsigned long long)__b, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_gt(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchdbs(__a, __b, &__cc);
+  return __cc == 0;
+}
+
+/*-- vec_all_le -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__a, (vector signed char)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs((vector signed char)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__a, (vector unsigned char)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__a,
+                        (vector unsigned char)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__a, (vector signed short)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs((vector signed short)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__a, (vector unsigned short)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__a,
+                        (vector unsigned short)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__a, (vector signed int)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs((vector signed int)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__a, (vector unsigned int)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__a,
+                        (vector unsigned int)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__a, (vector signed long long)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs((vector signed long long)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__a, (vector unsigned long long)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__a, __b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__a,
+                        (vector unsigned long long)__b, &__cc);
+  return __cc == 3;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_le(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchedbs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+/*-- vec_all_lt -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchbs((vector signed char)__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__b, (vector signed char)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__b, (vector unsigned char)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__b,
+                        (vector unsigned char)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchhs((vector signed short)__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__b, (vector signed short)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__b, (vector unsigned short)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__b,
+                        (vector unsigned short)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchfs((vector signed int)__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__b, (vector signed int)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__b, (vector unsigned int)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__b,
+                        (vector unsigned int)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchgs((vector signed long long)__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__b, (vector signed long long)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__b, __a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__b, (vector unsigned long long)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__b,
+                        (vector unsigned long long)__a, &__cc);
+  return __cc == 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_all_lt(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchdbs(__b, __a, &__cc);
+  return __cc == 0;
+}
+
+/*-- vec_all_nge ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_all_nge(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchedbs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+/*-- vec_all_ngt ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_all_ngt(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchdbs(__a, __b, &__cc);
+  return __cc == 3;
+}
+
+/*-- vec_all_nle ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_all_nle(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchedbs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+/*-- vec_all_nlt ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_all_nlt(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchdbs(__b, __a, &__cc);
+  return __cc == 3;
+}
+
+/*-- vec_all_nan ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_all_nan(vector double __a) {
+  int __cc;
+  __builtin_s390_vftcidb(__a, 15, &__cc);
+  return __cc == 0;
+}
+
+/*-- vec_all_numeric --------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_all_numeric(vector double __a) {
+  int __cc;
+  __builtin_s390_vftcidb(__a, 15, &__cc);
+  return __cc == 3;
+}
+
+/*-- vec_any_eq -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vceqbs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs(__a, (vector signed char)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vceqhs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs(__a, (vector signed short)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vceqfs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs(__a, (vector signed int)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs(__a, (vector signed long long)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_eq(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfcedbs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+/*-- vec_any_ne -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vceqbs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs(__a, (vector signed char)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vceqbs((vector signed char)__a,
+                        (vector signed char)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vceqhs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs(__a, (vector signed short)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vceqhs((vector signed short)__a,
+                        (vector signed short)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vceqfs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs(__a, (vector signed int)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vceqfs((vector signed int)__a,
+                        (vector signed int)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs(__a, (vector signed long long)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vceqgs((vector signed long long)__a,
+                        (vector signed long long)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ne(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfcedbs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+/*-- vec_any_ge -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchbs((vector signed char)__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__b, (vector signed char)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__b, (vector unsigned char)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__b,
+                        (vector unsigned char)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchhs((vector signed short)__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__b, (vector signed short)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__b, (vector unsigned short)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__b,
+                        (vector unsigned short)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchfs((vector signed int)__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__b, (vector signed int)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__b, (vector unsigned int)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__b,
+                        (vector unsigned int)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchgs((vector signed long long)__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__b, (vector signed long long)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__b, __a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__b, (vector unsigned long long)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__b,
+                        (vector unsigned long long)__a, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_ge(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchedbs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+/*-- vec_any_gt -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__a, (vector signed char)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs((vector signed char)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__a, (vector unsigned char)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__a,
+                        (vector unsigned char)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__a, (vector signed short)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs((vector signed short)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__a, (vector unsigned short)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__a,
+                        (vector unsigned short)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__a, (vector signed int)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs((vector signed int)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__a, (vector unsigned int)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__a,
+                        (vector unsigned int)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__a, (vector signed long long)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs((vector signed long long)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__a, (vector unsigned long long)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__a,
+                        (vector unsigned long long)__b, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_gt(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchdbs(__a, __b, &__cc);
+  return __cc <= 1;
+}
+
+/*-- vec_any_le -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__a, (vector signed char)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs((vector signed char)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__a, (vector unsigned char)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__a,
+                        (vector unsigned char)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__a, (vector signed short)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs((vector signed short)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__a, (vector unsigned short)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__a,
+                        (vector unsigned short)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__a, (vector signed int)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs((vector signed int)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__a, (vector unsigned int)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__a,
+                        (vector unsigned int)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__a, (vector signed long long)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs((vector signed long long)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__a, (vector unsigned long long)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__a, __b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__a,
+                        (vector unsigned long long)__b, &__cc);
+  return __cc != 0;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_le(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchedbs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+/*-- vec_any_lt -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector signed char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector signed char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchbs((vector signed char)__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool char __a, vector signed char __b) {
+  int __cc;
+  __builtin_s390_vchbs(__b, (vector signed char)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector unsigned char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector unsigned char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool char __a, vector unsigned char __b) {
+  int __cc;
+  __builtin_s390_vchlbs(__b, (vector unsigned char)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool char __a, vector bool char __b) {
+  int __cc;
+  __builtin_s390_vchlbs((vector unsigned char)__b,
+                        (vector unsigned char)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector signed short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector signed short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchhs((vector signed short)__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool short __a, vector signed short __b) {
+  int __cc;
+  __builtin_s390_vchhs(__b, (vector signed short)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector unsigned short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector unsigned short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool short __a, vector unsigned short __b) {
+  int __cc;
+  __builtin_s390_vchlhs(__b, (vector unsigned short)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool short __a, vector bool short __b) {
+  int __cc;
+  __builtin_s390_vchlhs((vector unsigned short)__b,
+                        (vector unsigned short)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector signed int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector signed int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchfs((vector signed int)__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool int __a, vector signed int __b) {
+  int __cc;
+  __builtin_s390_vchfs(__b, (vector signed int)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector unsigned int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector unsigned int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool int __a, vector unsigned int __b) {
+  int __cc;
+  __builtin_s390_vchlfs(__b, (vector unsigned int)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool int __a, vector bool int __b) {
+  int __cc;
+  __builtin_s390_vchlfs((vector unsigned int)__b,
+                        (vector unsigned int)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector signed long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector signed long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchgs((vector signed long long)__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool long long __a, vector signed long long __b) {
+  int __cc;
+  __builtin_s390_vchgs(__b, (vector signed long long)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector unsigned long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector unsigned long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool long long __a, vector unsigned long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs(__b, (vector unsigned long long)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector bool long long __a, vector bool long long __b) {
+  int __cc;
+  __builtin_s390_vchlgs((vector unsigned long long)__b,
+                        (vector unsigned long long)__a, &__cc);
+  return __cc <= 1;
+}
+
+static inline __ATTRS_o_ai int
+vec_any_lt(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchdbs(__b, __a, &__cc);
+  return __cc <= 1;
+}
+
+/*-- vec_any_nge ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_any_nge(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchedbs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+/*-- vec_any_ngt ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_any_ngt(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchdbs(__a, __b, &__cc);
+  return __cc != 0;
+}
+
+/*-- vec_any_nle ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_any_nle(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchedbs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+/*-- vec_any_nlt ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_any_nlt(vector double __a, vector double __b) {
+  int __cc;
+  __builtin_s390_vfchdbs(__b, __a, &__cc);
+  return __cc != 0;
+}
+
+/*-- vec_any_nan ------------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_any_nan(vector double __a) {
+  int __cc;
+  __builtin_s390_vftcidb(__a, 15, &__cc);
+  return __cc != 3;
+}
+
+/*-- vec_any_numeric --------------------------------------------------------*/
+
+static inline __ATTRS_ai int
+vec_any_numeric(vector double __a) {
+  int __cc;
+  __builtin_s390_vftcidb(__a, 15, &__cc);
+  return __cc != 0;
+}
+
+/*-- vec_andc ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_andc(vector bool char __a, vector bool char __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_andc(vector signed char __a, vector signed char __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_andc(vector bool char __a, vector signed char __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_andc(vector signed char __a, vector bool char __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_andc(vector unsigned char __a, vector unsigned char __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_andc(vector bool char __a, vector unsigned char __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_andc(vector unsigned char __a, vector bool char __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_andc(vector bool short __a, vector bool short __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_andc(vector signed short __a, vector signed short __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_andc(vector bool short __a, vector signed short __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_andc(vector signed short __a, vector bool short __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_andc(vector unsigned short __a, vector unsigned short __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_andc(vector bool short __a, vector unsigned short __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_andc(vector unsigned short __a, vector bool short __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_andc(vector bool int __a, vector bool int __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_andc(vector signed int __a, vector signed int __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_andc(vector bool int __a, vector signed int __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_andc(vector signed int __a, vector bool int __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_andc(vector unsigned int __a, vector unsigned int __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_andc(vector bool int __a, vector unsigned int __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_andc(vector unsigned int __a, vector bool int __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_andc(vector bool long long __a, vector bool long long __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_andc(vector signed long long __a, vector signed long long __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_andc(vector bool long long __a, vector signed long long __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_andc(vector signed long long __a, vector bool long long __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_andc(vector unsigned long long __a, vector unsigned long long __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_andc(vector bool long long __a, vector unsigned long long __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_andc(vector unsigned long long __a, vector bool long long __b) {
+  return __a & ~__b;
+}
+
+static inline __ATTRS_o_ai vector double
+vec_andc(vector double __a, vector double __b) {
+  return (vector double)((vector unsigned long long)__a &
+                         ~(vector unsigned long long)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_andc(vector bool long long __a, vector double __b) {
+  return (vector double)((vector unsigned long long)__a &
+                         ~(vector unsigned long long)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_andc(vector double __a, vector bool long long __b) {
+  return (vector double)((vector unsigned long long)__a &
+                         ~(vector unsigned long long)__b);
+}
+
+/*-- vec_nor ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_nor(vector bool char __a, vector bool char __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_nor(vector signed char __a, vector signed char __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_nor(vector bool char __a, vector signed char __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_nor(vector signed char __a, vector bool char __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_nor(vector unsigned char __a, vector unsigned char __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_nor(vector bool char __a, vector unsigned char __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_nor(vector unsigned char __a, vector bool char __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_nor(vector bool short __a, vector bool short __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_nor(vector signed short __a, vector signed short __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_nor(vector bool short __a, vector signed short __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_nor(vector signed short __a, vector bool short __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_nor(vector unsigned short __a, vector unsigned short __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_nor(vector bool short __a, vector unsigned short __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_nor(vector unsigned short __a, vector bool short __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_nor(vector bool int __a, vector bool int __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_nor(vector signed int __a, vector signed int __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_nor(vector bool int __a, vector signed int __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_nor(vector signed int __a, vector bool int __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_nor(vector unsigned int __a, vector unsigned int __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_nor(vector bool int __a, vector unsigned int __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_nor(vector unsigned int __a, vector bool int __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_nor(vector bool long long __a, vector bool long long __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_nor(vector signed long long __a, vector signed long long __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_nor(vector bool long long __a, vector signed long long __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_nor(vector signed long long __a, vector bool long long __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_nor(vector unsigned long long __a, vector unsigned long long __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_nor(vector bool long long __a, vector unsigned long long __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_nor(vector unsigned long long __a, vector bool long long __b) {
+  return ~(__a | __b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_nor(vector double __a, vector double __b) {
+  return (vector double)~((vector unsigned long long)__a |
+                          (vector unsigned long long)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_nor(vector bool long long __a, vector double __b) {
+  return (vector double)~((vector unsigned long long)__a |
+                          (vector unsigned long long)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_nor(vector double __a, vector bool long long __b) {
+  return (vector double)~((vector unsigned long long)__a |
+                          (vector unsigned long long)__b);
+}
+
+/*-- vec_cntlz --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cntlz(vector signed char __a) {
+  return __builtin_s390_vclzb((vector unsigned char)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cntlz(vector unsigned char __a) {
+  return __builtin_s390_vclzb(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cntlz(vector signed short __a) {
+  return __builtin_s390_vclzh((vector unsigned short)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cntlz(vector unsigned short __a) {
+  return __builtin_s390_vclzh(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cntlz(vector signed int __a) {
+  return __builtin_s390_vclzf((vector unsigned int)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cntlz(vector unsigned int __a) {
+  return __builtin_s390_vclzf(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_cntlz(vector signed long long __a) {
+  return __builtin_s390_vclzg((vector unsigned long long)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_cntlz(vector unsigned long long __a) {
+  return __builtin_s390_vclzg(__a);
+}
+
+/*-- vec_cnttz --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cnttz(vector signed char __a) {
+  return __builtin_s390_vctzb((vector unsigned char)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cnttz(vector unsigned char __a) {
+  return __builtin_s390_vctzb(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cnttz(vector signed short __a) {
+  return __builtin_s390_vctzh((vector unsigned short)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cnttz(vector unsigned short __a) {
+  return __builtin_s390_vctzh(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cnttz(vector signed int __a) {
+  return __builtin_s390_vctzf((vector unsigned int)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cnttz(vector unsigned int __a) {
+  return __builtin_s390_vctzf(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_cnttz(vector signed long long __a) {
+  return __builtin_s390_vctzg((vector unsigned long long)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_cnttz(vector unsigned long long __a) {
+  return __builtin_s390_vctzg(__a);
+}
+
+/*-- vec_popcnt -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_popcnt(vector signed char __a) {
+  return __builtin_s390_vpopctb((vector unsigned char)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_popcnt(vector unsigned char __a) {
+  return __builtin_s390_vpopctb(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_popcnt(vector signed short __a) {
+  return __builtin_s390_vpopcth((vector unsigned short)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_popcnt(vector unsigned short __a) {
+  return __builtin_s390_vpopcth(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_popcnt(vector signed int __a) {
+  return __builtin_s390_vpopctf((vector unsigned int)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_popcnt(vector unsigned int __a) {
+  return __builtin_s390_vpopctf(__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_popcnt(vector signed long long __a) {
+  return __builtin_s390_vpopctg((vector unsigned long long)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_popcnt(vector unsigned long long __a) {
+  return __builtin_s390_vpopctg(__a);
+}
+
+/*-- vec_rl -----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_rl(vector signed char __a, vector unsigned char __b) {
+  return (vector signed char)__builtin_s390_verllvb(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_rl(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_verllvb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_rl(vector signed short __a, vector unsigned short __b) {
+  return (vector signed short)__builtin_s390_verllvh(
+    (vector unsigned short)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_rl(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_verllvh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_rl(vector signed int __a, vector unsigned int __b) {
+  return (vector signed int)__builtin_s390_verllvf(
+    (vector unsigned int)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_rl(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_verllvf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_rl(vector signed long long __a, vector unsigned long long __b) {
+  return (vector signed long long)__builtin_s390_verllvg(
+    (vector unsigned long long)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_rl(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_verllvg(__a, __b);
+}
+
+/*-- vec_rli ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_rli(vector signed char __a, unsigned long __b) {
+  return (vector signed char)__builtin_s390_verllb(
+    (vector unsigned char)__a, (int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_rli(vector unsigned char __a, unsigned long __b) {
+  return __builtin_s390_verllb(__a, (int)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_rli(vector signed short __a, unsigned long __b) {
+  return (vector signed short)__builtin_s390_verllh(
+    (vector unsigned short)__a, (int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_rli(vector unsigned short __a, unsigned long __b) {
+  return __builtin_s390_verllh(__a, (int)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_rli(vector signed int __a, unsigned long __b) {
+  return (vector signed int)__builtin_s390_verllf(
+    (vector unsigned int)__a, (int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_rli(vector unsigned int __a, unsigned long __b) {
+  return __builtin_s390_verllf(__a, (int)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_rli(vector signed long long __a, unsigned long __b) {
+  return (vector signed long long)__builtin_s390_verllg(
+    (vector unsigned long long)__a, (int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_rli(vector unsigned long long __a, unsigned long __b) {
+  return __builtin_s390_verllg(__a, (int)__b);
+}
+
+/*-- vec_rl_mask ------------------------------------------------------------*/
+
+extern __ATTRS_o vector signed char
+vec_rl_mask(vector signed char __a, vector unsigned char __b,
+            unsigned char __c) __constant(__c);
+
+extern __ATTRS_o vector unsigned char
+vec_rl_mask(vector unsigned char __a, vector unsigned char __b,
+            unsigned char __c) __constant(__c);
+
+extern __ATTRS_o vector signed short
+vec_rl_mask(vector signed short __a, vector unsigned short __b,
+            unsigned char __c) __constant(__c);
+
+extern __ATTRS_o vector unsigned short
+vec_rl_mask(vector unsigned short __a, vector unsigned short __b,
+            unsigned char __c) __constant(__c);
+
+extern __ATTRS_o vector signed int
+vec_rl_mask(vector signed int __a, vector unsigned int __b,
+            unsigned char __c) __constant(__c);
+
+extern __ATTRS_o vector unsigned int
+vec_rl_mask(vector unsigned int __a, vector unsigned int __b,
+            unsigned char __c) __constant(__c);
+
+extern __ATTRS_o vector signed long long
+vec_rl_mask(vector signed long long __a, vector unsigned long long __b,
+            unsigned char __c) __constant(__c);
+
+extern __ATTRS_o vector unsigned long long
+vec_rl_mask(vector unsigned long long __a, vector unsigned long long __b,
+            unsigned char __c) __constant(__c);
+
+#define vec_rl_mask(X, Y, Z) ((__typeof__((vec_rl_mask)((X), (Y), (Z)))) \
+  __extension__ ({ \
+    vector unsigned char __res; \
+    vector unsigned char __x = (vector unsigned char)(X); \
+    vector unsigned char __y = (vector unsigned char)(Y); \
+    switch (sizeof ((X)[0])) { \
+    case 1: __res = (vector unsigned char) __builtin_s390_verimb( \
+             (vector unsigned char)__x, (vector unsigned char)__x, \
+             (vector unsigned char)__y, (Z)); break; \
+    case 2: __res = (vector unsigned char) __builtin_s390_verimh( \
+             (vector unsigned short)__x, (vector unsigned short)__x, \
+             (vector unsigned short)__y, (Z)); break; \
+    case 4: __res = (vector unsigned char) __builtin_s390_verimf( \
+             (vector unsigned int)__x, (vector unsigned int)__x, \
+             (vector unsigned int)__y, (Z)); break; \
+    default: __res = (vector unsigned char) __builtin_s390_verimg( \
+             (vector unsigned long long)__x, (vector unsigned long long)__x, \
+             (vector unsigned long long)__y, (Z)); break; \
+    } __res; }))
+
+/*-- vec_sll ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_sll(vector signed char __a, vector unsigned char __b) {
+  return (vector signed char)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_sll(vector signed char __a, vector unsigned short __b) {
+  return (vector signed char)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_sll(vector signed char __a, vector unsigned int __b) {
+  return (vector signed char)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_sll(vector bool char __a, vector unsigned char __b) {
+  return (vector bool char)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_sll(vector bool char __a, vector unsigned short __b) {
+  return (vector bool char)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_sll(vector bool char __a, vector unsigned int __b) {
+  return (vector bool char)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sll(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vsl(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sll(vector unsigned char __a, vector unsigned short __b) {
+  return __builtin_s390_vsl(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sll(vector unsigned char __a, vector unsigned int __b) {
+  return __builtin_s390_vsl(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_sll(vector signed short __a, vector unsigned char __b) {
+  return (vector signed short)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_sll(vector signed short __a, vector unsigned short __b) {
+  return (vector signed short)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_sll(vector signed short __a, vector unsigned int __b) {
+  return (vector signed short)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_sll(vector bool short __a, vector unsigned char __b) {
+  return (vector bool short)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_sll(vector bool short __a, vector unsigned short __b) {
+  return (vector bool short)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_sll(vector bool short __a, vector unsigned int __b) {
+  return (vector bool short)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_sll(vector unsigned short __a, vector unsigned char __b) {
+  return (vector unsigned short)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_sll(vector unsigned short __a, vector unsigned short __b) {
+  return (vector unsigned short)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_sll(vector unsigned short __a, vector unsigned int __b) {
+  return (vector unsigned short)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_sll(vector signed int __a, vector unsigned char __b) {
+  return (vector signed int)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_sll(vector signed int __a, vector unsigned short __b) {
+  return (vector signed int)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_sll(vector signed int __a, vector unsigned int __b) {
+  return (vector signed int)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_sll(vector bool int __a, vector unsigned char __b) {
+  return (vector bool int)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_sll(vector bool int __a, vector unsigned short __b) {
+  return (vector bool int)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_sll(vector bool int __a, vector unsigned int __b) {
+  return (vector bool int)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sll(vector unsigned int __a, vector unsigned char __b) {
+  return (vector unsigned int)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sll(vector unsigned int __a, vector unsigned short __b) {
+  return (vector unsigned int)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sll(vector unsigned int __a, vector unsigned int __b) {
+  return (vector unsigned int)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_sll(vector signed long long __a, vector unsigned char __b) {
+  return (vector signed long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_sll(vector signed long long __a, vector unsigned short __b) {
+  return (vector signed long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_sll(vector signed long long __a, vector unsigned int __b) {
+  return (vector signed long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_sll(vector bool long long __a, vector unsigned char __b) {
+  return (vector bool long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_sll(vector bool long long __a, vector unsigned short __b) {
+  return (vector bool long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_sll(vector bool long long __a, vector unsigned int __b) {
+  return (vector bool long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sll(vector unsigned long long __a, vector unsigned char __b) {
+  return (vector unsigned long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sll(vector unsigned long long __a, vector unsigned short __b) {
+  return (vector unsigned long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sll(vector unsigned long long __a, vector unsigned int __b) {
+  return (vector unsigned long long)__builtin_s390_vsl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+/*-- vec_slb ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_slb(vector signed char __a, vector signed char __b) {
+  return (vector signed char)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_slb(vector signed char __a, vector unsigned char __b) {
+  return (vector signed char)__builtin_s390_vslb(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_slb(vector unsigned char __a, vector signed char __b) {
+  return __builtin_s390_vslb(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_slb(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vslb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_slb(vector signed short __a, vector signed short __b) {
+  return (vector signed short)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_slb(vector signed short __a, vector unsigned short __b) {
+  return (vector signed short)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_slb(vector unsigned short __a, vector signed short __b) {
+  return (vector unsigned short)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_slb(vector unsigned short __a, vector unsigned short __b) {
+  return (vector unsigned short)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_slb(vector signed int __a, vector signed int __b) {
+  return (vector signed int)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_slb(vector signed int __a, vector unsigned int __b) {
+  return (vector signed int)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_slb(vector unsigned int __a, vector signed int __b) {
+  return (vector unsigned int)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_slb(vector unsigned int __a, vector unsigned int __b) {
+  return (vector unsigned int)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_slb(vector signed long long __a, vector signed long long __b) {
+  return (vector signed long long)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_slb(vector signed long long __a, vector unsigned long long __b) {
+  return (vector signed long long)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_slb(vector unsigned long long __a, vector signed long long __b) {
+  return (vector unsigned long long)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_slb(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector unsigned long long)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_slb(vector double __a, vector signed long long __b) {
+  return (vector double)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_slb(vector double __a, vector unsigned long long __b) {
+  return (vector double)__builtin_s390_vslb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+/*-- vec_sld ----------------------------------------------------------------*/
+
+extern __ATTRS_o vector signed char
+vec_sld(vector signed char __a, vector signed char __b, int __c)
+  __constant_range(__c, 0, 15);
+
+extern __ATTRS_o vector unsigned char
+vec_sld(vector unsigned char __a, vector unsigned char __b, int __c)
+  __constant_range(__c, 0, 15);
+
+extern __ATTRS_o vector signed short
+vec_sld(vector signed short __a, vector signed short __b, int __c)
+  __constant_range(__c, 0, 15);
+
+extern __ATTRS_o vector unsigned short
+vec_sld(vector unsigned short __a, vector unsigned short __b, int __c)
+  __constant_range(__c, 0, 15);
+
+extern __ATTRS_o vector signed int
+vec_sld(vector signed int __a, vector signed int __b, int __c)
+  __constant_range(__c, 0, 15);
+
+extern __ATTRS_o vector unsigned int
+vec_sld(vector unsigned int __a, vector unsigned int __b, int __c)
+  __constant_range(__c, 0, 15);
+
+extern __ATTRS_o vector signed long long
+vec_sld(vector signed long long __a, vector signed long long __b, int __c)
+  __constant_range(__c, 0, 15);
+
+extern __ATTRS_o vector unsigned long long
+vec_sld(vector unsigned long long __a, vector unsigned long long __b, int __c)
+  __constant_range(__c, 0, 15);
+
+extern __ATTRS_o vector double
+vec_sld(vector double __a, vector double __b, int __c)
+  __constant_range(__c, 0, 15);
+
+#define vec_sld(X, Y, Z) ((__typeof__((vec_sld)((X), (Y), (Z)))) \
+  __builtin_s390_vsldb((vector unsigned char)(X), \
+                       (vector unsigned char)(Y), (Z)))
+
+/*-- vec_sldw ---------------------------------------------------------------*/
+
+extern __ATTRS_o vector signed char
+vec_sldw(vector signed char __a, vector signed char __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector unsigned char
+vec_sldw(vector unsigned char __a, vector unsigned char __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector signed short
+vec_sldw(vector signed short __a, vector signed short __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector unsigned short
+vec_sldw(vector unsigned short __a, vector unsigned short __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector signed int
+vec_sldw(vector signed int __a, vector signed int __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector unsigned int
+vec_sldw(vector unsigned int __a, vector unsigned int __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector signed long long
+vec_sldw(vector signed long long __a, vector signed long long __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector unsigned long long
+vec_sldw(vector unsigned long long __a, vector unsigned long long __b, int __c)
+  __constant_range(__c, 0, 3);
+
+extern __ATTRS_o vector double
+vec_sldw(vector double __a, vector double __b, int __c)
+  __constant_range(__c, 0, 3);
+
+#define vec_sldw(X, Y, Z) ((__typeof__((vec_sldw)((X), (Y), (Z)))) \
+  __builtin_s390_vsldb((vector unsigned char)(X), \
+                       (vector unsigned char)(Y), (Z) * 4))
+
+/*-- vec_sral ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_sral(vector signed char __a, vector unsigned char __b) {
+  return (vector signed char)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_sral(vector signed char __a, vector unsigned short __b) {
+  return (vector signed char)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_sral(vector signed char __a, vector unsigned int __b) {
+  return (vector signed char)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_sral(vector bool char __a, vector unsigned char __b) {
+  return (vector bool char)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_sral(vector bool char __a, vector unsigned short __b) {
+  return (vector bool char)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_sral(vector bool char __a, vector unsigned int __b) {
+  return (vector bool char)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sral(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vsra(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sral(vector unsigned char __a, vector unsigned short __b) {
+  return __builtin_s390_vsra(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sral(vector unsigned char __a, vector unsigned int __b) {
+  return __builtin_s390_vsra(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_sral(vector signed short __a, vector unsigned char __b) {
+  return (vector signed short)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_sral(vector signed short __a, vector unsigned short __b) {
+  return (vector signed short)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_sral(vector signed short __a, vector unsigned int __b) {
+  return (vector signed short)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_sral(vector bool short __a, vector unsigned char __b) {
+  return (vector bool short)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_sral(vector bool short __a, vector unsigned short __b) {
+  return (vector bool short)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_sral(vector bool short __a, vector unsigned int __b) {
+  return (vector bool short)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_sral(vector unsigned short __a, vector unsigned char __b) {
+  return (vector unsigned short)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_sral(vector unsigned short __a, vector unsigned short __b) {
+  return (vector unsigned short)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_sral(vector unsigned short __a, vector unsigned int __b) {
+  return (vector unsigned short)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_sral(vector signed int __a, vector unsigned char __b) {
+  return (vector signed int)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_sral(vector signed int __a, vector unsigned short __b) {
+  return (vector signed int)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_sral(vector signed int __a, vector unsigned int __b) {
+  return (vector signed int)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_sral(vector bool int __a, vector unsigned char __b) {
+  return (vector bool int)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_sral(vector bool int __a, vector unsigned short __b) {
+  return (vector bool int)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_sral(vector bool int __a, vector unsigned int __b) {
+  return (vector bool int)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sral(vector unsigned int __a, vector unsigned char __b) {
+  return (vector unsigned int)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sral(vector unsigned int __a, vector unsigned short __b) {
+  return (vector unsigned int)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sral(vector unsigned int __a, vector unsigned int __b) {
+  return (vector unsigned int)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_sral(vector signed long long __a, vector unsigned char __b) {
+  return (vector signed long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_sral(vector signed long long __a, vector unsigned short __b) {
+  return (vector signed long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_sral(vector signed long long __a, vector unsigned int __b) {
+  return (vector signed long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_sral(vector bool long long __a, vector unsigned char __b) {
+  return (vector bool long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_sral(vector bool long long __a, vector unsigned short __b) {
+  return (vector bool long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_sral(vector bool long long __a, vector unsigned int __b) {
+  return (vector bool long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sral(vector unsigned long long __a, vector unsigned char __b) {
+  return (vector unsigned long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sral(vector unsigned long long __a, vector unsigned short __b) {
+  return (vector unsigned long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sral(vector unsigned long long __a, vector unsigned int __b) {
+  return (vector unsigned long long)__builtin_s390_vsra(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+/*-- vec_srab ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_srab(vector signed char __a, vector signed char __b) {
+  return (vector signed char)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_srab(vector signed char __a, vector unsigned char __b) {
+  return (vector signed char)__builtin_s390_vsrab(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_srab(vector unsigned char __a, vector signed char __b) {
+  return __builtin_s390_vsrab(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_srab(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vsrab(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_srab(vector signed short __a, vector signed short __b) {
+  return (vector signed short)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_srab(vector signed short __a, vector unsigned short __b) {
+  return (vector signed short)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_srab(vector unsigned short __a, vector signed short __b) {
+  return (vector unsigned short)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_srab(vector unsigned short __a, vector unsigned short __b) {
+  return (vector unsigned short)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_srab(vector signed int __a, vector signed int __b) {
+  return (vector signed int)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_srab(vector signed int __a, vector unsigned int __b) {
+  return (vector signed int)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_srab(vector unsigned int __a, vector signed int __b) {
+  return (vector unsigned int)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_srab(vector unsigned int __a, vector unsigned int __b) {
+  return (vector unsigned int)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_srab(vector signed long long __a, vector signed long long __b) {
+  return (vector signed long long)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_srab(vector signed long long __a, vector unsigned long long __b) {
+  return (vector signed long long)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_srab(vector unsigned long long __a, vector signed long long __b) {
+  return (vector unsigned long long)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_srab(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector unsigned long long)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_srab(vector double __a, vector signed long long __b) {
+  return (vector double)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_srab(vector double __a, vector unsigned long long __b) {
+  return (vector double)__builtin_s390_vsrab(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+/*-- vec_srl ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_srl(vector signed char __a, vector unsigned char __b) {
+  return (vector signed char)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_srl(vector signed char __a, vector unsigned short __b) {
+  return (vector signed char)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_srl(vector signed char __a, vector unsigned int __b) {
+  return (vector signed char)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_srl(vector bool char __a, vector unsigned char __b) {
+  return (vector bool char)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_srl(vector bool char __a, vector unsigned short __b) {
+  return (vector bool char)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_srl(vector bool char __a, vector unsigned int __b) {
+  return (vector bool char)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_srl(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vsrl(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_srl(vector unsigned char __a, vector unsigned short __b) {
+  return __builtin_s390_vsrl(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_srl(vector unsigned char __a, vector unsigned int __b) {
+  return __builtin_s390_vsrl(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_srl(vector signed short __a, vector unsigned char __b) {
+  return (vector signed short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_srl(vector signed short __a, vector unsigned short __b) {
+  return (vector signed short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_srl(vector signed short __a, vector unsigned int __b) {
+  return (vector signed short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_srl(vector bool short __a, vector unsigned char __b) {
+  return (vector bool short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_srl(vector bool short __a, vector unsigned short __b) {
+  return (vector bool short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_srl(vector bool short __a, vector unsigned int __b) {
+  return (vector bool short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_srl(vector unsigned short __a, vector unsigned char __b) {
+  return (vector unsigned short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_srl(vector unsigned short __a, vector unsigned short __b) {
+  return (vector unsigned short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_srl(vector unsigned short __a, vector unsigned int __b) {
+  return (vector unsigned short)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_srl(vector signed int __a, vector unsigned char __b) {
+  return (vector signed int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_srl(vector signed int __a, vector unsigned short __b) {
+  return (vector signed int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_srl(vector signed int __a, vector unsigned int __b) {
+  return (vector signed int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_srl(vector bool int __a, vector unsigned char __b) {
+  return (vector bool int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_srl(vector bool int __a, vector unsigned short __b) {
+  return (vector bool int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_srl(vector bool int __a, vector unsigned int __b) {
+  return (vector bool int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_srl(vector unsigned int __a, vector unsigned char __b) {
+  return (vector unsigned int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_srl(vector unsigned int __a, vector unsigned short __b) {
+  return (vector unsigned int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_srl(vector unsigned int __a, vector unsigned int __b) {
+  return (vector unsigned int)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_srl(vector signed long long __a, vector unsigned char __b) {
+  return (vector signed long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_srl(vector signed long long __a, vector unsigned short __b) {
+  return (vector signed long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_srl(vector signed long long __a, vector unsigned int __b) {
+  return (vector signed long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_srl(vector bool long long __a, vector unsigned char __b) {
+  return (vector bool long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_srl(vector bool long long __a, vector unsigned short __b) {
+  return (vector bool long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector bool long long
+vec_srl(vector bool long long __a, vector unsigned int __b) {
+  return (vector bool long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_srl(vector unsigned long long __a, vector unsigned char __b) {
+  return (vector unsigned long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_srl(vector unsigned long long __a, vector unsigned short __b) {
+  return (vector unsigned long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_srl(vector unsigned long long __a, vector unsigned int __b) {
+  return (vector unsigned long long)__builtin_s390_vsrl(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+/*-- vec_srb ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_srb(vector signed char __a, vector signed char __b) {
+  return (vector signed char)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_srb(vector signed char __a, vector unsigned char __b) {
+  return (vector signed char)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_srb(vector unsigned char __a, vector signed char __b) {
+  return __builtin_s390_vsrlb(__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_srb(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vsrlb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_srb(vector signed short __a, vector signed short __b) {
+  return (vector signed short)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_srb(vector signed short __a, vector unsigned short __b) {
+  return (vector signed short)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_srb(vector unsigned short __a, vector signed short __b) {
+  return (vector unsigned short)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_srb(vector unsigned short __a, vector unsigned short __b) {
+  return (vector unsigned short)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_srb(vector signed int __a, vector signed int __b) {
+  return (vector signed int)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_srb(vector signed int __a, vector unsigned int __b) {
+  return (vector signed int)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_srb(vector unsigned int __a, vector signed int __b) {
+  return (vector unsigned int)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_srb(vector unsigned int __a, vector unsigned int __b) {
+  return (vector unsigned int)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_srb(vector signed long long __a, vector signed long long __b) {
+  return (vector signed long long)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_srb(vector signed long long __a, vector unsigned long long __b) {
+  return (vector signed long long)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_srb(vector unsigned long long __a, vector signed long long __b) {
+  return (vector unsigned long long)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_srb(vector unsigned long long __a, vector unsigned long long __b) {
+  return (vector unsigned long long)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_srb(vector double __a, vector signed long long __b) {
+  return (vector double)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector double
+vec_srb(vector double __a, vector unsigned long long __b) {
+  return (vector double)__builtin_s390_vsrlb(
+    (vector unsigned char)__a, (vector unsigned char)__b);
+}
+
+/*-- vec_abs ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_abs(vector signed char __a) {
+  return vec_sel(__a, -__a, vec_cmplt(__a, (vector signed char)0));
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_abs(vector signed short __a) {
+  return vec_sel(__a, -__a, vec_cmplt(__a, (vector signed short)0));
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_abs(vector signed int __a) {
+  return vec_sel(__a, -__a, vec_cmplt(__a, (vector signed int)0));
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_abs(vector signed long long __a) {
+  return vec_sel(__a, -__a, vec_cmplt(__a, (vector signed long long)0));
+}
+
+static inline __ATTRS_o_ai vector double
+vec_abs(vector double __a) {
+  return __builtin_s390_vflpdb(__a);
+}
+
+/*-- vec_nabs ---------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_nabs(vector double __a) {
+  return __builtin_s390_vflndb(__a);
+}
+
+/*-- vec_max ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_max(vector signed char __a, vector signed char __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_max(vector signed char __a, vector bool char __b) {
+  vector signed char __bc = (vector signed char)__b;
+  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_max(vector bool char __a, vector signed char __b) {
+  vector signed char __ac = (vector signed char)__a;
+  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_max(vector unsigned char __a, vector unsigned char __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_max(vector unsigned char __a, vector bool char __b) {
+  vector unsigned char __bc = (vector unsigned char)__b;
+  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_max(vector bool char __a, vector unsigned char __b) {
+  vector unsigned char __ac = (vector unsigned char)__a;
+  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_max(vector signed short __a, vector signed short __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_max(vector signed short __a, vector bool short __b) {
+  vector signed short __bc = (vector signed short)__b;
+  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_max(vector bool short __a, vector signed short __b) {
+  vector signed short __ac = (vector signed short)__a;
+  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_max(vector unsigned short __a, vector unsigned short __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_max(vector unsigned short __a, vector bool short __b) {
+  vector unsigned short __bc = (vector unsigned short)__b;
+  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_max(vector bool short __a, vector unsigned short __b) {
+  vector unsigned short __ac = (vector unsigned short)__a;
+  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_max(vector signed int __a, vector signed int __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_max(vector signed int __a, vector bool int __b) {
+  vector signed int __bc = (vector signed int)__b;
+  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_max(vector bool int __a, vector signed int __b) {
+  vector signed int __ac = (vector signed int)__a;
+  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_max(vector unsigned int __a, vector unsigned int __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_max(vector unsigned int __a, vector bool int __b) {
+  vector unsigned int __bc = (vector unsigned int)__b;
+  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_max(vector bool int __a, vector unsigned int __b) {
+  vector unsigned int __ac = (vector unsigned int)__a;
+  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_max(vector signed long long __a, vector signed long long __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_max(vector signed long long __a, vector bool long long __b) {
+  vector signed long long __bc = (vector signed long long)__b;
+  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_max(vector bool long long __a, vector signed long long __b) {
+  vector signed long long __ac = (vector signed long long)__a;
+  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_max(vector unsigned long long __a, vector unsigned long long __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_max(vector unsigned long long __a, vector bool long long __b) {
+  vector unsigned long long __bc = (vector unsigned long long)__b;
+  return vec_sel(__bc, __a, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_max(vector bool long long __a, vector unsigned long long __b) {
+  vector unsigned long long __ac = (vector unsigned long long)__a;
+  return vec_sel(__b, __ac, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector double
+vec_max(vector double __a, vector double __b) {
+  return vec_sel(__b, __a, vec_cmpgt(__a, __b));
+}
+
+/*-- vec_min ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_min(vector signed char __a, vector signed char __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_min(vector signed char __a, vector bool char __b) {
+  vector signed char __bc = (vector signed char)__b;
+  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_min(vector bool char __a, vector signed char __b) {
+  vector signed char __ac = (vector signed char)__a;
+  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_min(vector unsigned char __a, vector unsigned char __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_min(vector unsigned char __a, vector bool char __b) {
+  vector unsigned char __bc = (vector unsigned char)__b;
+  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_min(vector bool char __a, vector unsigned char __b) {
+  vector unsigned char __ac = (vector unsigned char)__a;
+  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_min(vector signed short __a, vector signed short __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_min(vector signed short __a, vector bool short __b) {
+  vector signed short __bc = (vector signed short)__b;
+  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_min(vector bool short __a, vector signed short __b) {
+  vector signed short __ac = (vector signed short)__a;
+  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_min(vector unsigned short __a, vector unsigned short __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_min(vector unsigned short __a, vector bool short __b) {
+  vector unsigned short __bc = (vector unsigned short)__b;
+  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_min(vector bool short __a, vector unsigned short __b) {
+  vector unsigned short __ac = (vector unsigned short)__a;
+  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_min(vector signed int __a, vector signed int __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_min(vector signed int __a, vector bool int __b) {
+  vector signed int __bc = (vector signed int)__b;
+  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_min(vector bool int __a, vector signed int __b) {
+  vector signed int __ac = (vector signed int)__a;
+  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_min(vector unsigned int __a, vector unsigned int __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_min(vector unsigned int __a, vector bool int __b) {
+  vector unsigned int __bc = (vector unsigned int)__b;
+  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_min(vector bool int __a, vector unsigned int __b) {
+  vector unsigned int __ac = (vector unsigned int)__a;
+  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_min(vector signed long long __a, vector signed long long __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_min(vector signed long long __a, vector bool long long __b) {
+  vector signed long long __bc = (vector signed long long)__b;
+  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_min(vector bool long long __a, vector signed long long __b) {
+  vector signed long long __ac = (vector signed long long)__a;
+  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_min(vector unsigned long long __a, vector unsigned long long __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_min(vector unsigned long long __a, vector bool long long __b) {
+  vector unsigned long long __bc = (vector unsigned long long)__b;
+  return vec_sel(__a, __bc, vec_cmpgt(__a, __bc));
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_min(vector bool long long __a, vector unsigned long long __b) {
+  vector unsigned long long __ac = (vector unsigned long long)__a;
+  return vec_sel(__ac, __b, vec_cmpgt(__ac, __b));
+}
+
+static inline __ATTRS_o_ai vector double
+vec_min(vector double __a, vector double __b) {
+  return vec_sel(__a, __b, vec_cmpgt(__a, __b));
+}
+
+/*-- vec_add_u128 -----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_add_u128(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vaq(__a, __b);
+}
+
+/*-- vec_addc ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_addc(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vaccb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_addc(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vacch(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_addc(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vaccf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_addc(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vaccg(__a, __b);
+}
+
+/*-- vec_addc_u128 ----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_addc_u128(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vaccq(__a, __b);
+}
+
+/*-- vec_adde_u128 ----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_adde_u128(vector unsigned char __a, vector unsigned char __b,
+              vector unsigned char __c) {
+  return __builtin_s390_vacq(__a, __b, __c);
+}
+
+/*-- vec_addec_u128 ---------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_addec_u128(vector unsigned char __a, vector unsigned char __b,
+               vector unsigned char __c) {
+  return __builtin_s390_vacccq(__a, __b, __c);
+}
+
+/*-- vec_avg ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_avg(vector signed char __a, vector signed char __b) {
+  return __builtin_s390_vavgb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_avg(vector signed short __a, vector signed short __b) {
+  return __builtin_s390_vavgh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_avg(vector signed int __a, vector signed int __b) {
+  return __builtin_s390_vavgf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_avg(vector signed long long __a, vector signed long long __b) {
+  return __builtin_s390_vavgg(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_avg(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vavglb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_avg(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vavglh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_avg(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vavglf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_avg(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vavglg(__a, __b);
+}
+
+/*-- vec_checksum -----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned int
+vec_checksum(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vcksm(__a, __b);
+}
+
+/*-- vec_gfmsum -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_gfmsum(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vgfmb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_gfmsum(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vgfmh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_gfmsum(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vgfmf(__a, __b);
+}
+
+/*-- vec_gfmsum_128 ---------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_gfmsum_128(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vgfmg(__a, __b);
+}
+
+/*-- vec_gfmsum_accum -------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_gfmsum_accum(vector unsigned char __a, vector unsigned char __b,
+                 vector unsigned short __c) {
+  return __builtin_s390_vgfmab(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_gfmsum_accum(vector unsigned short __a, vector unsigned short __b,
+                 vector unsigned int __c) {
+  return __builtin_s390_vgfmah(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_gfmsum_accum(vector unsigned int __a, vector unsigned int __b,
+                 vector unsigned long long __c) {
+  return __builtin_s390_vgfmaf(__a, __b, __c);
+}
+
+/*-- vec_gfmsum_accum_128 ---------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_gfmsum_accum_128(vector unsigned long long __a,
+                     vector unsigned long long __b,
+                     vector unsigned char __c) {
+  return __builtin_s390_vgfmag(__a, __b, __c);
+}
+
+/*-- vec_mladd --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_mladd(vector signed char __a, vector signed char __b,
+          vector signed char __c) {
+  return __a * __b + __c;
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_mladd(vector unsigned char __a, vector signed char __b,
+          vector signed char __c) {
+  return (vector signed char)__a * __b + __c;
+}
+
+static inline __ATTRS_o_ai vector signed char
+vec_mladd(vector signed char __a, vector unsigned char __b,
+          vector unsigned char __c) {
+  return __a * (vector signed char)__b + (vector signed char)__c;
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_mladd(vector unsigned char __a, vector unsigned char __b,
+          vector unsigned char __c) {
+  return __a * __b + __c;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_mladd(vector signed short __a, vector signed short __b,
+          vector signed short __c) {
+  return __a * __b + __c;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_mladd(vector unsigned short __a, vector signed short __b,
+          vector signed short __c) {
+  return (vector signed short)__a * __b + __c;
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_mladd(vector signed short __a, vector unsigned short __b,
+          vector unsigned short __c) {
+  return __a * (vector signed short)__b + (vector signed short)__c;
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_mladd(vector unsigned short __a, vector unsigned short __b,
+          vector unsigned short __c) {
+  return __a * __b + __c;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mladd(vector signed int __a, vector signed int __b,
+          vector signed int __c) {
+  return __a * __b + __c;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mladd(vector unsigned int __a, vector signed int __b,
+          vector signed int __c) {
+  return (vector signed int)__a * __b + __c;
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mladd(vector signed int __a, vector unsigned int __b,
+          vector unsigned int __c) {
+  return __a * (vector signed int)__b + (vector signed int)__c;
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_mladd(vector unsigned int __a, vector unsigned int __b,
+          vector unsigned int __c) {
+  return __a * __b + __c;
+}
+
+/*-- vec_mhadd --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_mhadd(vector signed char __a, vector signed char __b,
+          vector signed char __c) {
+  return __builtin_s390_vmahb(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_mhadd(vector unsigned char __a, vector unsigned char __b,
+          vector unsigned char __c) {
+  return __builtin_s390_vmalhb(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_mhadd(vector signed short __a, vector signed short __b,
+          vector signed short __c) {
+  return __builtin_s390_vmahh(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_mhadd(vector unsigned short __a, vector unsigned short __b,
+          vector unsigned short __c) {
+  return __builtin_s390_vmalhh(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mhadd(vector signed int __a, vector signed int __b,
+          vector signed int __c) {
+  return __builtin_s390_vmahf(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_mhadd(vector unsigned int __a, vector unsigned int __b,
+          vector unsigned int __c) {
+  return __builtin_s390_vmalhf(__a, __b, __c);
+}
+
+/*-- vec_meadd --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed short
+vec_meadd(vector signed char __a, vector signed char __b,
+          vector signed short __c) {
+  return __builtin_s390_vmaeb(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_meadd(vector unsigned char __a, vector unsigned char __b,
+          vector unsigned short __c) {
+  return __builtin_s390_vmaleb(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_meadd(vector signed short __a, vector signed short __b,
+          vector signed int __c) {
+  return __builtin_s390_vmaeh(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_meadd(vector unsigned short __a, vector unsigned short __b,
+          vector unsigned int __c) {
+  return __builtin_s390_vmaleh(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_meadd(vector signed int __a, vector signed int __b,
+          vector signed long long __c) {
+  return __builtin_s390_vmaef(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_meadd(vector unsigned int __a, vector unsigned int __b,
+          vector unsigned long long __c) {
+  return __builtin_s390_vmalef(__a, __b, __c);
+}
+
+/*-- vec_moadd --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed short
+vec_moadd(vector signed char __a, vector signed char __b,
+          vector signed short __c) {
+  return __builtin_s390_vmaob(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_moadd(vector unsigned char __a, vector unsigned char __b,
+          vector unsigned short __c) {
+  return __builtin_s390_vmalob(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_moadd(vector signed short __a, vector signed short __b,
+          vector signed int __c) {
+  return __builtin_s390_vmaoh(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_moadd(vector unsigned short __a, vector unsigned short __b,
+          vector unsigned int __c) {
+  return __builtin_s390_vmaloh(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_moadd(vector signed int __a, vector signed int __b,
+          vector signed long long __c) {
+  return __builtin_s390_vmaof(__a, __b, __c);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_moadd(vector unsigned int __a, vector unsigned int __b,
+          vector unsigned long long __c) {
+  return __builtin_s390_vmalof(__a, __b, __c);
+}
+
+/*-- vec_mulh ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_mulh(vector signed char __a, vector signed char __b) {
+  return __builtin_s390_vmhb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_mulh(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vmlhb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_mulh(vector signed short __a, vector signed short __b) {
+  return __builtin_s390_vmhh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_mulh(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vmlhh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mulh(vector signed int __a, vector signed int __b) {
+  return __builtin_s390_vmhf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_mulh(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vmlhf(__a, __b);
+}
+
+/*-- vec_mule ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed short
+vec_mule(vector signed char __a, vector signed char __b) {
+  return __builtin_s390_vmeb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_mule(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vmleb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mule(vector signed short __a, vector signed short __b) {
+  return __builtin_s390_vmeh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_mule(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vmleh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_mule(vector signed int __a, vector signed int __b) {
+  return __builtin_s390_vmef(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_mule(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vmlef(__a, __b);
+}
+
+/*-- vec_mulo ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed short
+vec_mulo(vector signed char __a, vector signed char __b) {
+  return __builtin_s390_vmob(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_mulo(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vmlob(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_mulo(vector signed short __a, vector signed short __b) {
+  return __builtin_s390_vmoh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_mulo(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vmloh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed long long
+vec_mulo(vector signed int __a, vector signed int __b) {
+  return __builtin_s390_vmof(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_mulo(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vmlof(__a, __b);
+}
+
+/*-- vec_sub_u128 -----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_sub_u128(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vsq(__a, __b);
+}
+
+/*-- vec_subc ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_subc(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vscbib(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_subc(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vscbih(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_subc(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vscbif(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_subc(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vscbig(__a, __b);
+}
+
+/*-- vec_subc_u128 ----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_subc_u128(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vscbiq(__a, __b);
+}
+
+/*-- vec_sube_u128 ----------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_sube_u128(vector unsigned char __a, vector unsigned char __b,
+              vector unsigned char __c) {
+  return __builtin_s390_vsbiq(__a, __b, __c);
+}
+
+/*-- vec_subec_u128 ---------------------------------------------------------*/
+
+static inline __ATTRS_ai vector unsigned char
+vec_subec_u128(vector unsigned char __a, vector unsigned char __b,
+               vector unsigned char __c) {
+  return __builtin_s390_vsbcbiq(__a, __b, __c);
+}
+
+/*-- vec_sum2 ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sum2(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vsumgh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_sum2(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vsumgf(__a, __b);
+}
+
+/*-- vec_sum_u128 -----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sum_u128(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vsumqf(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_sum_u128(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vsumqg(__a, __b);
+}
+
+/*-- vec_sum4 ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sum4(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vsumb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_sum4(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vsumh(__a, __b);
+}
+
+/*-- vec_test_mask ----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector signed char __a, vector unsigned char __b) {
+  return __builtin_s390_vtm((vector unsigned char)__a,
+                            (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vtm(__a, __b);
+}
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector signed short __a, vector unsigned short __b) {
+  return __builtin_s390_vtm((vector unsigned char)__a,
+                            (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vtm((vector unsigned char)__a,
+                            (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector signed int __a, vector unsigned int __b) {
+  return __builtin_s390_vtm((vector unsigned char)__a,
+                            (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vtm((vector unsigned char)__a,
+                            (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector signed long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vtm((vector unsigned char)__a,
+                            (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector unsigned long long __a, vector unsigned long long __b) {
+  return __builtin_s390_vtm((vector unsigned char)__a,
+                            (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai int
+vec_test_mask(vector double __a, vector unsigned long long __b) {
+  return __builtin_s390_vtm((vector unsigned char)__a,
+                            (vector unsigned char)__b);
+}
+
+/*-- vec_madd ---------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_madd(vector double __a, vector double __b, vector double __c) {
+  return __builtin_s390_vfmadb(__a, __b, __c);
+}
+
+/*-- vec_msub ---------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_msub(vector double __a, vector double __b, vector double __c) {
+  return __builtin_s390_vfmsdb(__a, __b, __c);
+}
+
+/*-- vec_sqrt ---------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_sqrt(vector double __a) {
+  return __builtin_s390_vfsqdb(__a);
+}
+
+/*-- vec_ld2f ---------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_ld2f(const float *__ptr) {
+  typedef float __v2f32 __attribute__((__vector_size__(8)));
+  return __builtin_convertvector(*(const __v2f32 *)__ptr, vector double);
+}
+
+/*-- vec_st2f ---------------------------------------------------------------*/
+
+static inline __ATTRS_ai void
+vec_st2f(vector double __a, float *__ptr) {
+  typedef float __v2f32 __attribute__((__vector_size__(8)));
+  *(__v2f32 *)__ptr = __builtin_convertvector(__a, __v2f32);
+}
+
+/*-- vec_ctd ----------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector double
+vec_ctd(vector signed long long __a, int __b)
+  __constant_range(__b, 0, 31) {
+  vector double __conv = __builtin_convertvector(__a, vector double);
+  __conv *= (vector double)(vector unsigned long long)((0x3ffULL - __b) << 52);
+  return __conv;
+}
+
+static inline __ATTRS_o_ai vector double
+vec_ctd(vector unsigned long long __a, int __b)
+  __constant_range(__b, 0, 31) {
+  vector double __conv = __builtin_convertvector(__a, vector double);
+  __conv *= (vector double)(vector unsigned long long)((0x3ffULL - __b) << 52);
+  return __conv;
+}
+
+/*-- vec_ctsl ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed long long
+vec_ctsl(vector double __a, int __b)
+  __constant_range(__b, 0, 31) {
+  __a *= (vector double)(vector unsigned long long)((0x3ffULL + __b) << 52);
+  return __builtin_convertvector(__a, vector signed long long);
+}
+
+/*-- vec_ctul ---------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned long long
+vec_ctul(vector double __a, int __b)
+  __constant_range(__b, 0, 31) {
+  __a *= (vector double)(vector unsigned long long)((0x3ffULL + __b) << 52);
+  return __builtin_convertvector(__a, vector unsigned long long);
+}
+
+/*-- vec_roundp -------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_roundp(vector double __a) {
+  return __builtin_s390_vfidb(__a, 4, 6);
+}
+
+/*-- vec_ceil ---------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_ceil(vector double __a) {
+  // On this platform, vec_ceil never triggers the IEEE-inexact exception.
+  return __builtin_s390_vfidb(__a, 4, 6);
+}
+
+/*-- vec_roundm -------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_roundm(vector double __a) {
+  return __builtin_s390_vfidb(__a, 4, 7);
+}
+
+/*-- vec_floor --------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_floor(vector double __a) {
+  // On this platform, vec_floor never triggers the IEEE-inexact exception.
+  return __builtin_s390_vfidb(__a, 4, 7);
+}
+
+/*-- vec_roundz -------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_roundz(vector double __a) {
+  return __builtin_s390_vfidb(__a, 4, 5);
+}
+
+/*-- vec_trunc --------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_trunc(vector double __a) {
+  // On this platform, vec_trunc never triggers the IEEE-inexact exception.
+  return __builtin_s390_vfidb(__a, 4, 5);
+}
+
+/*-- vec_roundc -------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_roundc(vector double __a) {
+  return __builtin_s390_vfidb(__a, 4, 0);
+}
+
+/*-- vec_round --------------------------------------------------------------*/
+
+static inline __ATTRS_ai vector double
+vec_round(vector double __a) {
+  return __builtin_s390_vfidb(__a, 4, 4);
+}
+
+/*-- vec_fp_test_data_class -------------------------------------------------*/
+
+#define vec_fp_test_data_class(X, Y, Z) \
+  ((vector bool long long)__builtin_s390_vftcidb((X), (Y), (Z)))
+
+/*-- vec_cp_until_zero ------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cp_until_zero(vector signed char __a) {
+  return (vector signed char)__builtin_s390_vistrb((vector unsigned char)__a);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_cp_until_zero(vector bool char __a) {
+  return (vector bool char)__builtin_s390_vistrb((vector unsigned char)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cp_until_zero(vector unsigned char __a) {
+  return __builtin_s390_vistrb(__a);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cp_until_zero(vector signed short __a) {
+  return (vector signed short)__builtin_s390_vistrh((vector unsigned short)__a);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cp_until_zero(vector bool short __a) {
+  return (vector bool short)__builtin_s390_vistrh((vector unsigned short)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cp_until_zero(vector unsigned short __a) {
+  return __builtin_s390_vistrh(__a);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cp_until_zero(vector signed int __a) {
+  return (vector signed int)__builtin_s390_vistrf((vector unsigned int)__a);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cp_until_zero(vector bool int __a) {
+  return (vector bool int)__builtin_s390_vistrf((vector unsigned int)__a);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cp_until_zero(vector unsigned int __a) {
+  return __builtin_s390_vistrf(__a);
+}
+
+/*-- vec_cp_until_zero_cc ---------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cp_until_zero_cc(vector signed char __a, int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vistrbs((vector unsigned char)__a, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_cp_until_zero_cc(vector bool char __a, int *__cc) {
+  return (vector bool char)
+    __builtin_s390_vistrbs((vector unsigned char)__a, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cp_until_zero_cc(vector unsigned char __a, int *__cc) {
+  return __builtin_s390_vistrbs(__a, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cp_until_zero_cc(vector signed short __a, int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vistrhs((vector unsigned short)__a, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cp_until_zero_cc(vector bool short __a, int *__cc) {
+  return (vector bool short)
+    __builtin_s390_vistrhs((vector unsigned short)__a, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cp_until_zero_cc(vector unsigned short __a, int *__cc) {
+  return __builtin_s390_vistrhs(__a, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cp_until_zero_cc(vector signed int __a, int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vistrfs((vector unsigned int)__a, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cp_until_zero_cc(vector bool int __a, int *__cc) {
+  return (vector bool int)__builtin_s390_vistrfs((vector unsigned int)__a,
+                                                 __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cp_until_zero_cc(vector unsigned int __a, int *__cc) {
+  return __builtin_s390_vistrfs(__a, __cc);
+}
+
+/*-- vec_cmpeq_idx ----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cmpeq_idx(vector signed char __a, vector signed char __b) {
+  return (vector signed char)
+    __builtin_s390_vfeeb((vector unsigned char)__a,
+                         (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpeq_idx(vector bool char __a, vector bool char __b) {
+  return __builtin_s390_vfeeb((vector unsigned char)__a,
+                              (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpeq_idx(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vfeeb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cmpeq_idx(vector signed short __a, vector signed short __b) {
+  return (vector signed short)
+    __builtin_s390_vfeeh((vector unsigned short)__a,
+                         (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpeq_idx(vector bool short __a, vector bool short __b) {
+  return __builtin_s390_vfeeh((vector unsigned short)__a,
+                              (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpeq_idx(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vfeeh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cmpeq_idx(vector signed int __a, vector signed int __b) {
+  return (vector signed int)
+    __builtin_s390_vfeef((vector unsigned int)__a,
+                         (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpeq_idx(vector bool int __a, vector bool int __b) {
+  return __builtin_s390_vfeef((vector unsigned int)__a,
+                              (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpeq_idx(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vfeef(__a, __b);
+}
+
+/*-- vec_cmpeq_idx_cc -------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cmpeq_idx_cc(vector signed char __a, vector signed char __b, int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vfeebs((vector unsigned char)__a,
+                          (vector unsigned char)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpeq_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
+  return __builtin_s390_vfeebs((vector unsigned char)__a,
+                               (vector unsigned char)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpeq_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                 int *__cc) {
+  return __builtin_s390_vfeebs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cmpeq_idx_cc(vector signed short __a, vector signed short __b, int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vfeehs((vector unsigned short)__a,
+                          (vector unsigned short)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpeq_idx_cc(vector bool short __a, vector bool short __b, int *__cc) {
+  return __builtin_s390_vfeehs((vector unsigned short)__a,
+                               (vector unsigned short)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpeq_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                 int *__cc) {
+  return __builtin_s390_vfeehs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cmpeq_idx_cc(vector signed int __a, vector signed int __b, int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vfeefs((vector unsigned int)__a,
+                          (vector unsigned int)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpeq_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
+  return __builtin_s390_vfeefs((vector unsigned int)__a,
+                               (vector unsigned int)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpeq_idx_cc(vector unsigned int __a, vector unsigned int __b, int *__cc) {
+  return __builtin_s390_vfeefs(__a, __b, __cc);
+}
+
+/*-- vec_cmpeq_or_0_idx -----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cmpeq_or_0_idx(vector signed char __a, vector signed char __b) {
+  return (vector signed char)
+    __builtin_s390_vfeezb((vector unsigned char)__a,
+                          (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpeq_or_0_idx(vector bool char __a, vector bool char __b) {
+  return __builtin_s390_vfeezb((vector unsigned char)__a,
+                               (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpeq_or_0_idx(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vfeezb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cmpeq_or_0_idx(vector signed short __a, vector signed short __b) {
+  return (vector signed short)
+    __builtin_s390_vfeezh((vector unsigned short)__a,
+                          (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpeq_or_0_idx(vector bool short __a, vector bool short __b) {
+  return __builtin_s390_vfeezh((vector unsigned short)__a,
+                               (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpeq_or_0_idx(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vfeezh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cmpeq_or_0_idx(vector signed int __a, vector signed int __b) {
+  return (vector signed int)
+    __builtin_s390_vfeezf((vector unsigned int)__a,
+                          (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpeq_or_0_idx(vector bool int __a, vector bool int __b) {
+  return __builtin_s390_vfeezf((vector unsigned int)__a,
+                               (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpeq_or_0_idx(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vfeezf(__a, __b);
+}
+
+/*-- vec_cmpeq_or_0_idx_cc --------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cmpeq_or_0_idx_cc(vector signed char __a, vector signed char __b,
+                      int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vfeezbs((vector unsigned char)__a,
+                           (vector unsigned char)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpeq_or_0_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
+  return __builtin_s390_vfeezbs((vector unsigned char)__a,
+                                (vector unsigned char)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpeq_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                      int *__cc) {
+  return __builtin_s390_vfeezbs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cmpeq_or_0_idx_cc(vector signed short __a, vector signed short __b,
+                      int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vfeezhs((vector unsigned short)__a,
+                           (vector unsigned short)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpeq_or_0_idx_cc(vector bool short __a, vector bool short __b, int *__cc) {
+  return __builtin_s390_vfeezhs((vector unsigned short)__a,
+                                (vector unsigned short)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpeq_or_0_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                      int *__cc) {
+  return __builtin_s390_vfeezhs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cmpeq_or_0_idx_cc(vector signed int __a, vector signed int __b, int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vfeezfs((vector unsigned int)__a,
+                           (vector unsigned int)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpeq_or_0_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
+  return __builtin_s390_vfeezfs((vector unsigned int)__a,
+                                (vector unsigned int)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpeq_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                      int *__cc) {
+  return __builtin_s390_vfeezfs(__a, __b, __cc);
+}
+
+/*-- vec_cmpne_idx ----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cmpne_idx(vector signed char __a, vector signed char __b) {
+  return (vector signed char)
+    __builtin_s390_vfeneb((vector unsigned char)__a,
+                          (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpne_idx(vector bool char __a, vector bool char __b) {
+  return __builtin_s390_vfeneb((vector unsigned char)__a,
+                               (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpne_idx(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vfeneb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cmpne_idx(vector signed short __a, vector signed short __b) {
+  return (vector signed short)
+    __builtin_s390_vfeneh((vector unsigned short)__a,
+                          (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpne_idx(vector bool short __a, vector bool short __b) {
+  return __builtin_s390_vfeneh((vector unsigned short)__a,
+                               (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpne_idx(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vfeneh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cmpne_idx(vector signed int __a, vector signed int __b) {
+  return (vector signed int)
+    __builtin_s390_vfenef((vector unsigned int)__a,
+                          (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpne_idx(vector bool int __a, vector bool int __b) {
+  return __builtin_s390_vfenef((vector unsigned int)__a,
+                               (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpne_idx(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vfenef(__a, __b);
+}
+
+/*-- vec_cmpne_idx_cc -------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cmpne_idx_cc(vector signed char __a, vector signed char __b, int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vfenebs((vector unsigned char)__a,
+                           (vector unsigned char)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpne_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
+  return __builtin_s390_vfenebs((vector unsigned char)__a,
+                                (vector unsigned char)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpne_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                 int *__cc) {
+  return __builtin_s390_vfenebs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cmpne_idx_cc(vector signed short __a, vector signed short __b, int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vfenehs((vector unsigned short)__a,
+                           (vector unsigned short)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpne_idx_cc(vector bool short __a, vector bool short __b, int *__cc) {
+  return __builtin_s390_vfenehs((vector unsigned short)__a,
+                                (vector unsigned short)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpne_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                 int *__cc) {
+  return __builtin_s390_vfenehs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cmpne_idx_cc(vector signed int __a, vector signed int __b, int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vfenefs((vector unsigned int)__a,
+                           (vector unsigned int)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpne_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
+  return __builtin_s390_vfenefs((vector unsigned int)__a,
+                                (vector unsigned int)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpne_idx_cc(vector unsigned int __a, vector unsigned int __b, int *__cc) {
+  return __builtin_s390_vfenefs(__a, __b, __cc);
+}
+
+/*-- vec_cmpne_or_0_idx -----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cmpne_or_0_idx(vector signed char __a, vector signed char __b) {
+  return (vector signed char)
+    __builtin_s390_vfenezb((vector unsigned char)__a,
+                           (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpne_or_0_idx(vector bool char __a, vector bool char __b) {
+  return __builtin_s390_vfenezb((vector unsigned char)__a,
+                                (vector unsigned char)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpne_or_0_idx(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vfenezb(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cmpne_or_0_idx(vector signed short __a, vector signed short __b) {
+  return (vector signed short)
+    __builtin_s390_vfenezh((vector unsigned short)__a,
+                           (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpne_or_0_idx(vector bool short __a, vector bool short __b) {
+  return __builtin_s390_vfenezh((vector unsigned short)__a,
+                                (vector unsigned short)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpne_or_0_idx(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vfenezh(__a, __b);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cmpne_or_0_idx(vector signed int __a, vector signed int __b) {
+  return (vector signed int)
+    __builtin_s390_vfenezf((vector unsigned int)__a,
+                           (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpne_or_0_idx(vector bool int __a, vector bool int __b) {
+  return __builtin_s390_vfenezf((vector unsigned int)__a,
+                                (vector unsigned int)__b);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpne_or_0_idx(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vfenezf(__a, __b);
+}
+
+/*-- vec_cmpne_or_0_idx_cc --------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_cmpne_or_0_idx_cc(vector signed char __a, vector signed char __b,
+                      int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vfenezbs((vector unsigned char)__a,
+                            (vector unsigned char)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpne_or_0_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
+  return __builtin_s390_vfenezbs((vector unsigned char)__a,
+                                 (vector unsigned char)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpne_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                      int *__cc) {
+  return __builtin_s390_vfenezbs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_cmpne_or_0_idx_cc(vector signed short __a, vector signed short __b,
+                      int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vfenezhs((vector unsigned short)__a,
+                            (vector unsigned short)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpne_or_0_idx_cc(vector bool short __a, vector bool short __b, int *__cc) {
+  return __builtin_s390_vfenezhs((vector unsigned short)__a,
+                                 (vector unsigned short)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpne_or_0_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                      int *__cc) {
+  return __builtin_s390_vfenezhs(__a, __b, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_cmpne_or_0_idx_cc(vector signed int __a, vector signed int __b, int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vfenezfs((vector unsigned int)__a,
+                            (vector unsigned int)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpne_or_0_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
+  return __builtin_s390_vfenezfs((vector unsigned int)__a,
+                                 (vector unsigned int)__b, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpne_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                      int *__cc) {
+  return __builtin_s390_vfenezfs(__a, __b, __cc);
+}
+
+/*-- vec_cmprg --------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmprg(vector unsigned char __a, vector unsigned char __b,
+          vector unsigned char __c) {
+  return (vector bool char)__builtin_s390_vstrcb(__a, __b, __c, 4);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmprg(vector unsigned short __a, vector unsigned short __b,
+          vector unsigned short __c) {
+  return (vector bool short)__builtin_s390_vstrch(__a, __b, __c, 4);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmprg(vector unsigned int __a, vector unsigned int __b,
+          vector unsigned int __c) {
+  return (vector bool int)__builtin_s390_vstrcf(__a, __b, __c, 4);
+}
+
+/*-- vec_cmprg_cc -----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmprg_cc(vector unsigned char __a, vector unsigned char __b,
+             vector unsigned char __c, int *__cc) {
+  return (vector bool char)__builtin_s390_vstrcbs(__a, __b, __c, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmprg_cc(vector unsigned short __a, vector unsigned short __b,
+             vector unsigned short __c, int *__cc) {
+  return (vector bool short)__builtin_s390_vstrchs(__a, __b, __c, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmprg_cc(vector unsigned int __a, vector unsigned int __b,
+             vector unsigned int __c, int *__cc) {
+  return (vector bool int)__builtin_s390_vstrcfs(__a, __b, __c, 4, __cc);
+}
+
+/*-- vec_cmprg_idx ----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmprg_idx(vector unsigned char __a, vector unsigned char __b,
+              vector unsigned char __c) {
+  return __builtin_s390_vstrcb(__a, __b, __c, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmprg_idx(vector unsigned short __a, vector unsigned short __b,
+              vector unsigned short __c) {
+  return __builtin_s390_vstrch(__a, __b, __c, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmprg_idx(vector unsigned int __a, vector unsigned int __b,
+              vector unsigned int __c) {
+  return __builtin_s390_vstrcf(__a, __b, __c, 0);
+}
+
+/*-- vec_cmprg_idx_cc -------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmprg_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                 vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrcbs(__a, __b, __c, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmprg_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                 vector unsigned short __c, int *__cc) {
+  return __builtin_s390_vstrchs(__a, __b, __c, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmprg_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                 vector unsigned int __c, int *__cc) {
+  return __builtin_s390_vstrcfs(__a, __b, __c, 0, __cc);
+}
+
+/*-- vec_cmprg_or_0_idx -----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmprg_or_0_idx(vector unsigned char __a, vector unsigned char __b,
+                   vector unsigned char __c) {
+  return __builtin_s390_vstrczb(__a, __b, __c, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmprg_or_0_idx(vector unsigned short __a, vector unsigned short __b,
+                   vector unsigned short __c) {
+  return __builtin_s390_vstrczh(__a, __b, __c, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmprg_or_0_idx(vector unsigned int __a, vector unsigned int __b,
+                   vector unsigned int __c) {
+  return __builtin_s390_vstrczf(__a, __b, __c, 0);
+}
+
+/*-- vec_cmprg_or_0_idx_cc --------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmprg_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                      vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrczbs(__a, __b, __c, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmprg_or_0_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                      vector unsigned short __c, int *__cc) {
+  return __builtin_s390_vstrczhs(__a, __b, __c, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmprg_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                      vector unsigned int __c, int *__cc) {
+  return __builtin_s390_vstrczfs(__a, __b, __c, 0, __cc);
+}
+
+/*-- vec_cmpnrg -------------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpnrg(vector unsigned char __a, vector unsigned char __b,
+           vector unsigned char __c) {
+  return (vector bool char)__builtin_s390_vstrcb(__a, __b, __c, 12);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpnrg(vector unsigned short __a, vector unsigned short __b,
+           vector unsigned short __c) {
+  return (vector bool short)__builtin_s390_vstrch(__a, __b, __c, 12);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpnrg(vector unsigned int __a, vector unsigned int __b,
+           vector unsigned int __c) {
+  return (vector bool int)__builtin_s390_vstrcf(__a, __b, __c, 12);
+}
+
+/*-- vec_cmpnrg_cc ----------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_cmpnrg_cc(vector unsigned char __a, vector unsigned char __b,
+              vector unsigned char __c, int *__cc) {
+  return (vector bool char)__builtin_s390_vstrcbs(__a, __b, __c, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_cmpnrg_cc(vector unsigned short __a, vector unsigned short __b,
+              vector unsigned short __c, int *__cc) {
+  return (vector bool short)__builtin_s390_vstrchs(__a, __b, __c, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_cmpnrg_cc(vector unsigned int __a, vector unsigned int __b,
+              vector unsigned int __c, int *__cc) {
+  return (vector bool int)__builtin_s390_vstrcfs(__a, __b, __c, 12, __cc);
+}
+
+/*-- vec_cmpnrg_idx ---------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpnrg_idx(vector unsigned char __a, vector unsigned char __b,
+               vector unsigned char __c) {
+  return __builtin_s390_vstrcb(__a, __b, __c, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpnrg_idx(vector unsigned short __a, vector unsigned short __b,
+               vector unsigned short __c) {
+  return __builtin_s390_vstrch(__a, __b, __c, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpnrg_idx(vector unsigned int __a, vector unsigned int __b,
+               vector unsigned int __c) {
+  return __builtin_s390_vstrcf(__a, __b, __c, 8);
+}
+
+/*-- vec_cmpnrg_idx_cc ------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpnrg_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                  vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrcbs(__a, __b, __c, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpnrg_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                  vector unsigned short __c, int *__cc) {
+  return __builtin_s390_vstrchs(__a, __b, __c, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpnrg_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                  vector unsigned int __c, int *__cc) {
+  return __builtin_s390_vstrcfs(__a, __b, __c, 8, __cc);
+}
+
+/*-- vec_cmpnrg_or_0_idx ----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpnrg_or_0_idx(vector unsigned char __a, vector unsigned char __b,
+                    vector unsigned char __c) {
+  return __builtin_s390_vstrczb(__a, __b, __c, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpnrg_or_0_idx(vector unsigned short __a, vector unsigned short __b,
+                    vector unsigned short __c) {
+  return __builtin_s390_vstrczh(__a, __b, __c, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpnrg_or_0_idx(vector unsigned int __a, vector unsigned int __b,
+                    vector unsigned int __c) {
+  return __builtin_s390_vstrczf(__a, __b, __c, 8);
+}
+
+/*-- vec_cmpnrg_or_0_idx_cc -------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_cmpnrg_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                       vector unsigned char __c, int *__cc) {
+  return __builtin_s390_vstrczbs(__a, __b, __c, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_cmpnrg_or_0_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                       vector unsigned short __c, int *__cc) {
+  return __builtin_s390_vstrczhs(__a, __b, __c, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_cmpnrg_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                       vector unsigned int __c, int *__cc) {
+  return __builtin_s390_vstrczfs(__a, __b, __c, 8, __cc);
+}
+
+/*-- vec_find_any_eq --------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_eq(vector signed char __a, vector signed char __b) {
+  return (vector bool char)
+    __builtin_s390_vfaeb((vector unsigned char)__a,
+                         (vector unsigned char)__b, 4);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_eq(vector bool char __a, vector bool char __b) {
+  return (vector bool char)
+    __builtin_s390_vfaeb((vector unsigned char)__a,
+                         (vector unsigned char)__b, 4);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_eq(vector unsigned char __a, vector unsigned char __b) {
+  return (vector bool char)__builtin_s390_vfaeb(__a, __b, 4);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_eq(vector signed short __a, vector signed short __b) {
+  return (vector bool short)
+    __builtin_s390_vfaeh((vector unsigned short)__a,
+                         (vector unsigned short)__b, 4);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_eq(vector bool short __a, vector bool short __b) {
+  return (vector bool short)
+    __builtin_s390_vfaeh((vector unsigned short)__a,
+                         (vector unsigned short)__b, 4);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_eq(vector unsigned short __a, vector unsigned short __b) {
+  return (vector bool short)__builtin_s390_vfaeh(__a, __b, 4);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_eq(vector signed int __a, vector signed int __b) {
+  return (vector bool int)
+    __builtin_s390_vfaef((vector unsigned int)__a,
+                         (vector unsigned int)__b, 4);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_eq(vector bool int __a, vector bool int __b) {
+  return (vector bool int)
+    __builtin_s390_vfaef((vector unsigned int)__a,
+                         (vector unsigned int)__b, 4);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_eq(vector unsigned int __a, vector unsigned int __b) {
+  return (vector bool int)__builtin_s390_vfaef(__a, __b, 4);
+}
+
+/*-- vec_find_any_eq_cc -----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_eq_cc(vector signed char __a, vector signed char __b, int *__cc) {
+  return (vector bool char)
+    __builtin_s390_vfaebs((vector unsigned char)__a,
+                          (vector unsigned char)__b, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_eq_cc(vector bool char __a, vector bool char __b, int *__cc) {
+  return (vector bool char)
+    __builtin_s390_vfaebs((vector unsigned char)__a,
+                          (vector unsigned char)__b, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_eq_cc(vector unsigned char __a, vector unsigned char __b,
+                   int *__cc) {
+  return (vector bool char)__builtin_s390_vfaebs(__a, __b, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_eq_cc(vector signed short __a, vector signed short __b,
+                   int *__cc) {
+  return (vector bool short)
+    __builtin_s390_vfaehs((vector unsigned short)__a,
+                          (vector unsigned short)__b, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_eq_cc(vector bool short __a, vector bool short __b, int *__cc) {
+  return (vector bool short)
+    __builtin_s390_vfaehs((vector unsigned short)__a,
+                          (vector unsigned short)__b, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_eq_cc(vector unsigned short __a, vector unsigned short __b,
+                   int *__cc) {
+  return (vector bool short)__builtin_s390_vfaehs(__a, __b, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_eq_cc(vector signed int __a, vector signed int __b, int *__cc) {
+  return (vector bool int)
+    __builtin_s390_vfaefs((vector unsigned int)__a,
+                          (vector unsigned int)__b, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_eq_cc(vector bool int __a, vector bool int __b, int *__cc) {
+  return (vector bool int)
+    __builtin_s390_vfaefs((vector unsigned int)__a,
+                          (vector unsigned int)__b, 4, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_eq_cc(vector unsigned int __a, vector unsigned int __b,
+                   int *__cc) {
+  return (vector bool int)__builtin_s390_vfaefs(__a, __b, 4, __cc);
+}
+
+/*-- vec_find_any_eq_idx ----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_find_any_eq_idx(vector signed char __a, vector signed char __b) {
+  return (vector signed char)
+    __builtin_s390_vfaeb((vector unsigned char)__a,
+                         (vector unsigned char)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_eq_idx(vector bool char __a, vector bool char __b) {
+  return __builtin_s390_vfaeb((vector unsigned char)__a,
+                              (vector unsigned char)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_eq_idx(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vfaeb(__a, __b, 0);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_find_any_eq_idx(vector signed short __a, vector signed short __b) {
+  return (vector signed short)
+    __builtin_s390_vfaeh((vector unsigned short)__a,
+                         (vector unsigned short)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_eq_idx(vector bool short __a, vector bool short __b) {
+  return __builtin_s390_vfaeh((vector unsigned short)__a,
+                              (vector unsigned short)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_eq_idx(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vfaeh(__a, __b, 0);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_find_any_eq_idx(vector signed int __a, vector signed int __b) {
+  return (vector signed int)
+    __builtin_s390_vfaef((vector unsigned int)__a,
+                         (vector unsigned int)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_eq_idx(vector bool int __a, vector bool int __b) {
+  return __builtin_s390_vfaef((vector unsigned int)__a,
+                              (vector unsigned int)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_eq_idx(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vfaef(__a, __b, 0);
+}
+
+/*-- vec_find_any_eq_idx_cc -------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_find_any_eq_idx_cc(vector signed char __a, vector signed char __b,
+                       int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vfaebs((vector unsigned char)__a,
+                          (vector unsigned char)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_eq_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
+  return __builtin_s390_vfaebs((vector unsigned char)__a,
+                               (vector unsigned char)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_eq_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                       int *__cc) {
+  return __builtin_s390_vfaebs(__a, __b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_find_any_eq_idx_cc(vector signed short __a, vector signed short __b,
+                       int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vfaehs((vector unsigned short)__a,
+                          (vector unsigned short)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_eq_idx_cc(vector bool short __a, vector bool short __b,
+                       int *__cc) {
+  return __builtin_s390_vfaehs((vector unsigned short)__a,
+                               (vector unsigned short)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_eq_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                       int *__cc) {
+  return __builtin_s390_vfaehs(__a, __b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_find_any_eq_idx_cc(vector signed int __a, vector signed int __b,
+                       int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vfaefs((vector unsigned int)__a,
+                          (vector unsigned int)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_eq_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
+  return __builtin_s390_vfaefs((vector unsigned int)__a,
+                               (vector unsigned int)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_eq_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                       int *__cc) {
+  return __builtin_s390_vfaefs(__a, __b, 0, __cc);
+}
+
+/*-- vec_find_any_eq_or_0_idx -----------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_find_any_eq_or_0_idx(vector signed char __a, vector signed char __b) {
+  return (vector signed char)
+    __builtin_s390_vfaezb((vector unsigned char)__a,
+                          (vector unsigned char)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_eq_or_0_idx(vector bool char __a, vector bool char __b) {
+  return __builtin_s390_vfaezb((vector unsigned char)__a,
+                               (vector unsigned char)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_eq_or_0_idx(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vfaezb(__a, __b, 0);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_find_any_eq_or_0_idx(vector signed short __a, vector signed short __b) {
+  return (vector signed short)
+    __builtin_s390_vfaezh((vector unsigned short)__a,
+                          (vector unsigned short)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_eq_or_0_idx(vector bool short __a, vector bool short __b) {
+  return __builtin_s390_vfaezh((vector unsigned short)__a,
+                               (vector unsigned short)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_eq_or_0_idx(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vfaezh(__a, __b, 0);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_find_any_eq_or_0_idx(vector signed int __a, vector signed int __b) {
+  return (vector signed int)
+    __builtin_s390_vfaezf((vector unsigned int)__a,
+                          (vector unsigned int)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_eq_or_0_idx(vector bool int __a, vector bool int __b) {
+  return __builtin_s390_vfaezf((vector unsigned int)__a,
+                               (vector unsigned int)__b, 0);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_eq_or_0_idx(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vfaezf(__a, __b, 0);
+}
+
+/*-- vec_find_any_eq_or_0_idx_cc --------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_find_any_eq_or_0_idx_cc(vector signed char __a, vector signed char __b,
+                            int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vfaezbs((vector unsigned char)__a,
+                           (vector unsigned char)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_eq_or_0_idx_cc(vector bool char __a, vector bool char __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezbs((vector unsigned char)__a,
+                                (vector unsigned char)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_eq_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezbs(__a, __b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_find_any_eq_or_0_idx_cc(vector signed short __a, vector signed short __b,
+                            int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vfaezhs((vector unsigned short)__a,
+                           (vector unsigned short)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_eq_or_0_idx_cc(vector bool short __a, vector bool short __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezhs((vector unsigned short)__a,
+                                (vector unsigned short)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_eq_or_0_idx_cc(vector unsigned short __a,
+                            vector unsigned short __b, int *__cc) {
+  return __builtin_s390_vfaezhs(__a, __b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_find_any_eq_or_0_idx_cc(vector signed int __a, vector signed int __b,
+                            int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vfaezfs((vector unsigned int)__a,
+                           (vector unsigned int)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_eq_or_0_idx_cc(vector bool int __a, vector bool int __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezfs((vector unsigned int)__a,
+                                (vector unsigned int)__b, 0, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_eq_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezfs(__a, __b, 0, __cc);
+}
+
+/*-- vec_find_any_ne --------------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_ne(vector signed char __a, vector signed char __b) {
+  return (vector bool char)
+    __builtin_s390_vfaeb((vector unsigned char)__a,
+                         (vector unsigned char)__b, 12);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_ne(vector bool char __a, vector bool char __b) {
+  return (vector bool char)
+    __builtin_s390_vfaeb((vector unsigned char)__a,
+                         (vector unsigned char)__b, 12);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_ne(vector unsigned char __a, vector unsigned char __b) {
+  return (vector bool char)__builtin_s390_vfaeb(__a, __b, 12);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_ne(vector signed short __a, vector signed short __b) {
+  return (vector bool short)
+    __builtin_s390_vfaeh((vector unsigned short)__a,
+                         (vector unsigned short)__b, 12);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_ne(vector bool short __a, vector bool short __b) {
+  return (vector bool short)
+    __builtin_s390_vfaeh((vector unsigned short)__a,
+                         (vector unsigned short)__b, 12);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_ne(vector unsigned short __a, vector unsigned short __b) {
+  return (vector bool short)__builtin_s390_vfaeh(__a, __b, 12);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_ne(vector signed int __a, vector signed int __b) {
+  return (vector bool int)
+    __builtin_s390_vfaef((vector unsigned int)__a,
+                         (vector unsigned int)__b, 12);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_ne(vector bool int __a, vector bool int __b) {
+  return (vector bool int)
+    __builtin_s390_vfaef((vector unsigned int)__a,
+                         (vector unsigned int)__b, 12);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_ne(vector unsigned int __a, vector unsigned int __b) {
+  return (vector bool int)__builtin_s390_vfaef(__a, __b, 12);
+}
+
+/*-- vec_find_any_ne_cc -----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_ne_cc(vector signed char __a, vector signed char __b, int *__cc) {
+  return (vector bool char)
+    __builtin_s390_vfaebs((vector unsigned char)__a,
+                          (vector unsigned char)__b, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_ne_cc(vector bool char __a, vector bool char __b, int *__cc) {
+  return (vector bool char)
+    __builtin_s390_vfaebs((vector unsigned char)__a,
+                          (vector unsigned char)__b, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool char
+vec_find_any_ne_cc(vector unsigned char __a, vector unsigned char __b,
+                   int *__cc) {
+  return (vector bool char)__builtin_s390_vfaebs(__a, __b, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_ne_cc(vector signed short __a, vector signed short __b,
+                   int *__cc) {
+  return (vector bool short)
+    __builtin_s390_vfaehs((vector unsigned short)__a,
+                          (vector unsigned short)__b, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_ne_cc(vector bool short __a, vector bool short __b, int *__cc) {
+  return (vector bool short)
+    __builtin_s390_vfaehs((vector unsigned short)__a,
+                          (vector unsigned short)__b, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool short
+vec_find_any_ne_cc(vector unsigned short __a, vector unsigned short __b,
+                   int *__cc) {
+  return (vector bool short)__builtin_s390_vfaehs(__a, __b, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_ne_cc(vector signed int __a, vector signed int __b, int *__cc) {
+  return (vector bool int)
+    __builtin_s390_vfaefs((vector unsigned int)__a,
+                          (vector unsigned int)__b, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_ne_cc(vector bool int __a, vector bool int __b, int *__cc) {
+  return (vector bool int)
+    __builtin_s390_vfaefs((vector unsigned int)__a,
+                          (vector unsigned int)__b, 12, __cc);
+}
+
+static inline __ATTRS_o_ai vector bool int
+vec_find_any_ne_cc(vector unsigned int __a, vector unsigned int __b,
+                   int *__cc) {
+  return (vector bool int)__builtin_s390_vfaefs(__a, __b, 12, __cc);
+}
+
+/*-- vec_find_any_ne_idx ----------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_find_any_ne_idx(vector signed char __a, vector signed char __b) {
+  return (vector signed char)
+    __builtin_s390_vfaeb((vector unsigned char)__a,
+                         (vector unsigned char)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_ne_idx(vector bool char __a, vector bool char __b) {
+  return __builtin_s390_vfaeb((vector unsigned char)__a,
+                              (vector unsigned char)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_ne_idx(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vfaeb(__a, __b, 8);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_find_any_ne_idx(vector signed short __a, vector signed short __b) {
+  return (vector signed short)
+    __builtin_s390_vfaeh((vector unsigned short)__a,
+                         (vector unsigned short)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_ne_idx(vector bool short __a, vector bool short __b) {
+  return __builtin_s390_vfaeh((vector unsigned short)__a,
+                              (vector unsigned short)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_ne_idx(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vfaeh(__a, __b, 8);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_find_any_ne_idx(vector signed int __a, vector signed int __b) {
+  return (vector signed int)
+    __builtin_s390_vfaef((vector unsigned int)__a,
+                         (vector unsigned int)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_ne_idx(vector bool int __a, vector bool int __b) {
+  return __builtin_s390_vfaef((vector unsigned int)__a,
+                              (vector unsigned int)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_ne_idx(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vfaef(__a, __b, 8);
+}
+
+/*-- vec_find_any_ne_idx_cc -------------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_find_any_ne_idx_cc(vector signed char __a, vector signed char __b,
+                       int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vfaebs((vector unsigned char)__a,
+                          (vector unsigned char)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_ne_idx_cc(vector bool char __a, vector bool char __b, int *__cc) {
+  return __builtin_s390_vfaebs((vector unsigned char)__a,
+                               (vector unsigned char)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_ne_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                       int *__cc) {
+  return __builtin_s390_vfaebs(__a, __b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_find_any_ne_idx_cc(vector signed short __a, vector signed short __b,
+                       int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vfaehs((vector unsigned short)__a,
+                          (vector unsigned short)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_ne_idx_cc(vector bool short __a, vector bool short __b,
+                       int *__cc) {
+  return __builtin_s390_vfaehs((vector unsigned short)__a,
+                               (vector unsigned short)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_ne_idx_cc(vector unsigned short __a, vector unsigned short __b,
+                       int *__cc) {
+  return __builtin_s390_vfaehs(__a, __b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_find_any_ne_idx_cc(vector signed int __a, vector signed int __b,
+                       int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vfaefs((vector unsigned int)__a,
+                          (vector unsigned int)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_ne_idx_cc(vector bool int __a, vector bool int __b, int *__cc) {
+  return __builtin_s390_vfaefs((vector unsigned int)__a,
+                               (vector unsigned int)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_ne_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                       int *__cc) {
+  return __builtin_s390_vfaefs(__a, __b, 8, __cc);
+}
+
+/*-- vec_find_any_ne_or_0_idx -----------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_find_any_ne_or_0_idx(vector signed char __a, vector signed char __b) {
+  return (vector signed char)
+    __builtin_s390_vfaezb((vector unsigned char)__a,
+                          (vector unsigned char)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_ne_or_0_idx(vector bool char __a, vector bool char __b) {
+  return __builtin_s390_vfaezb((vector unsigned char)__a,
+                               (vector unsigned char)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_ne_or_0_idx(vector unsigned char __a, vector unsigned char __b) {
+  return __builtin_s390_vfaezb(__a, __b, 8);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_find_any_ne_or_0_idx(vector signed short __a, vector signed short __b) {
+  return (vector signed short)
+    __builtin_s390_vfaezh((vector unsigned short)__a,
+                          (vector unsigned short)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_ne_or_0_idx(vector bool short __a, vector bool short __b) {
+  return __builtin_s390_vfaezh((vector unsigned short)__a,
+                               (vector unsigned short)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_ne_or_0_idx(vector unsigned short __a, vector unsigned short __b) {
+  return __builtin_s390_vfaezh(__a, __b, 8);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_find_any_ne_or_0_idx(vector signed int __a, vector signed int __b) {
+  return (vector signed int)
+    __builtin_s390_vfaezf((vector unsigned int)__a,
+                          (vector unsigned int)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_ne_or_0_idx(vector bool int __a, vector bool int __b) {
+  return __builtin_s390_vfaezf((vector unsigned int)__a,
+                               (vector unsigned int)__b, 8);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_ne_or_0_idx(vector unsigned int __a, vector unsigned int __b) {
+  return __builtin_s390_vfaezf(__a, __b, 8);
+}
+
+/*-- vec_find_any_ne_or_0_idx_cc --------------------------------------------*/
+
+static inline __ATTRS_o_ai vector signed char
+vec_find_any_ne_or_0_idx_cc(vector signed char __a, vector signed char __b,
+                            int *__cc) {
+  return (vector signed char)
+    __builtin_s390_vfaezbs((vector unsigned char)__a,
+                           (vector unsigned char)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_ne_or_0_idx_cc(vector bool char __a, vector bool char __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezbs((vector unsigned char)__a,
+                                (vector unsigned char)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned char
+vec_find_any_ne_or_0_idx_cc(vector unsigned char __a, vector unsigned char __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezbs(__a, __b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed short
+vec_find_any_ne_or_0_idx_cc(vector signed short __a, vector signed short __b,
+                            int *__cc) {
+  return (vector signed short)
+    __builtin_s390_vfaezhs((vector unsigned short)__a,
+                           (vector unsigned short)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_ne_or_0_idx_cc(vector bool short __a, vector bool short __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezhs((vector unsigned short)__a,
+                                (vector unsigned short)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned short
+vec_find_any_ne_or_0_idx_cc(vector unsigned short __a,
+                            vector unsigned short __b, int *__cc) {
+  return __builtin_s390_vfaezhs(__a, __b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector signed int
+vec_find_any_ne_or_0_idx_cc(vector signed int __a, vector signed int __b,
+                            int *__cc) {
+  return (vector signed int)
+    __builtin_s390_vfaezfs((vector unsigned int)__a,
+                           (vector unsigned int)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_ne_or_0_idx_cc(vector bool int __a, vector bool int __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezfs((vector unsigned int)__a,
+                                (vector unsigned int)__b, 8, __cc);
+}
+
+static inline __ATTRS_o_ai vector unsigned int
+vec_find_any_ne_or_0_idx_cc(vector unsigned int __a, vector unsigned int __b,
+                            int *__cc) {
+  return __builtin_s390_vfaezfs(__a, __b, 8, __cc);
+}
+
+#undef __constant_pow2_range
+#undef __constant_range
+#undef __constant
+#undef __ATTRS_o
+#undef __ATTRS_o_ai
+#undef __ATTRS_ai
+
+#else
+
+#error "Use -fzvector to enable vector extensions"
+
+#endif
diff --git a/clang/lib/Parse/Parser.cpp b/clang/lib/Parse/Parser.cpp
index e76f767..0574a63 100644
--- a/clang/lib/Parse/Parser.cpp
+++ b/clang/lib/Parse/Parser.cpp
@@ -476,11 +476,15 @@
 
   Ident_super = &PP.getIdentifierTable().get("super");
 
-  if (getLangOpts().AltiVec) {
+  Ident_vector = nullptr;
+  Ident_bool = nullptr;
+  Ident_pixel = nullptr;
+  if (getLangOpts().AltiVec || getLangOpts().ZVector) {
     Ident_vector = &PP.getIdentifierTable().get("vector");
-    Ident_pixel = &PP.getIdentifierTable().get("pixel");
     Ident_bool = &PP.getIdentifierTable().get("bool");
   }
+  if (getLangOpts().AltiVec)
+    Ident_pixel = &PP.getIdentifierTable().get("pixel");
 
   Ident_introduced = nullptr;
   Ident_deprecated = nullptr;
diff --git a/clang/lib/Sema/DeclSpec.cpp b/clang/lib/Sema/DeclSpec.cpp
index ea3872f..4adbb2b 100644
--- a/clang/lib/Sema/DeclSpec.cpp
+++ b/clang/lib/Sema/DeclSpec.cpp
@@ -987,10 +987,11 @@
         Diag(D, TSWLoc, diag::err_invalid_vector_bool_decl_spec)
           << getSpecifierName((TSW)TypeSpecWidth);
 
-      // vector bool long long requires VSX support.
+      // vector bool long long requires VSX support or ZVector.
       if ((TypeSpecWidth == TSW_longlong) &&
           (!PP.getTargetInfo().hasFeature("vsx")) &&
-          (!PP.getTargetInfo().hasFeature("power8-vector")))
+          (!PP.getTargetInfo().hasFeature("power8-vector")) &&
+          !PP.getLangOpts().ZVector)
         Diag(D, TSTLoc, diag::err_invalid_vector_long_long_decl_spec);
 
       // Elements of vector bool are interpreted as unsigned. (PIM 2.1)
@@ -999,14 +1000,23 @@
         TypeSpecSign = TSS_unsigned;
     } else if (TypeSpecType == TST_double) {
       // vector long double and vector long long double are never allowed.
-      // vector double is OK for Power7 and later.
+      // vector double is OK for Power7 and later, and ZVector.
       if (TypeSpecWidth == TSW_long || TypeSpecWidth == TSW_longlong)
         Diag(D, TSWLoc, diag::err_invalid_vector_long_double_decl_spec);
-      else if (!PP.getTargetInfo().hasFeature("vsx"))
+      else if (!PP.getTargetInfo().hasFeature("vsx") &&
+               !PP.getLangOpts().ZVector)
         Diag(D, TSTLoc, diag::err_invalid_vector_double_decl_spec);
+    } else if (TypeSpecType == TST_float) {
+      // vector float is unsupported for ZVector.
+      if (PP.getLangOpts().ZVector)
+        Diag(D, TSTLoc, diag::err_invalid_vector_float_decl_spec);
     } else if (TypeSpecWidth == TSW_long) {
-      Diag(D, TSWLoc, diag::warn_vector_long_decl_spec_combination)
-        << getSpecifierName((TST)TypeSpecType, Policy);
+      // vector long is unsupported for ZVector and deprecated for AltiVec.
+      if (PP.getLangOpts().ZVector)
+        Diag(D, TSWLoc, diag::err_invalid_vector_long_decl_spec);
+      else
+        Diag(D, TSWLoc, diag::warn_vector_long_decl_spec_combination)
+          << getSpecifierName((TST)TypeSpecType, Policy);
     }
 
     if (TypeAltiVecPixel) {
diff --git a/clang/lib/Sema/SemaExpr.cpp b/clang/lib/Sema/SemaExpr.cpp
index 1ae983c..617b06a 100644
--- a/clang/lib/Sema/SemaExpr.cpp
+++ b/clang/lib/Sema/SemaExpr.cpp
@@ -5518,7 +5518,7 @@
   // i.e. all the elements are integer constants.
   ParenExpr *PE = dyn_cast<ParenExpr>(CastExpr);
   ParenListExpr *PLE = dyn_cast<ParenListExpr>(CastExpr);
-  if ((getLangOpts().AltiVec || getLangOpts().OpenCL)
+  if ((getLangOpts().AltiVec || getLangOpts().ZVector || getLangOpts().OpenCL)
        && castType->isVectorType() && (PE || PLE)) {
     if (PLE && PLE->getNumExprs() == 0) {
       Diag(PLE->getExprLoc(), diag::err_altivec_empty_initializer);
@@ -6075,7 +6075,9 @@
   if (LHS.get()->getType()->isVectorType() ||
       RHS.get()->getType()->isVectorType()) {
     QualType VecResTy = S.CheckVectorOperands(LHS, RHS, QuestionLoc,
-                                              /*isCompAssign*/false);
+                                              /*isCompAssign*/false,
+                                              /*AllowBothBool*/true,
+                                              /*AllowBoolConversions*/false);
     if (VecResTy.isNull()) return QualType();
     // The result type must match the condition type as specified in
     // OpenCL v1.1 s6.11.6.
@@ -6126,7 +6128,9 @@
   // Now check the two expressions.
   if (LHS.get()->getType()->isVectorType() ||
       RHS.get()->getType()->isVectorType())
-    return CheckVectorOperands(LHS, RHS, QuestionLoc, /*isCompAssign*/false);
+    return CheckVectorOperands(LHS, RHS, QuestionLoc, /*isCompAssign*/false,
+                               /*AllowBothBool*/true,
+                               /*AllowBoolConversions*/false);
 
   QualType ResTy = UsualArithmeticConversions(LHS, RHS);
   if (LHS.isInvalid() || RHS.isInvalid())
@@ -7267,7 +7271,9 @@
 }
 
 QualType Sema::CheckVectorOperands(ExprResult &LHS, ExprResult &RHS,
-                                   SourceLocation Loc, bool IsCompAssign) {
+                                   SourceLocation Loc, bool IsCompAssign,
+                                   bool AllowBothBool,
+                                   bool AllowBoolConversions) {
   if (!IsCompAssign) {
     LHS = DefaultFunctionArrayLvalueConversion(LHS.get());
     if (LHS.isInvalid())
@@ -7282,14 +7288,21 @@
   QualType LHSType = LHS.get()->getType().getUnqualifiedType();
   QualType RHSType = RHS.get()->getType().getUnqualifiedType();
 
-  // If the vector types are identical, return.
-  if (Context.hasSameType(LHSType, RHSType))
-    return LHSType;
-
   const VectorType *LHSVecType = LHSType->getAs<VectorType>();
   const VectorType *RHSVecType = RHSType->getAs<VectorType>();
   assert(LHSVecType || RHSVecType);
 
+  // AltiVec-style "vector bool op vector bool" combinations are allowed
+  // for some operators but not others.
+  if (!AllowBothBool &&
+      LHSVecType && LHSVecType->getVectorKind() == VectorType::AltiVecBool &&
+      RHSVecType && RHSVecType->getVectorKind() == VectorType::AltiVecBool)
+    return InvalidOperands(Loc, LHS, RHS);
+
+  // If the vector types are identical, return.
+  if (Context.hasSameType(LHSType, RHSType))
+    return LHSType;
+
   // If we have compatible AltiVec and GCC vector types, use the AltiVec type.
   if (LHSVecType && RHSVecType &&
       Context.areCompatibleVectorTypes(LHSType, RHSType)) {
@@ -7303,6 +7316,28 @@
     return RHSType;
   }
 
+  // AllowBoolConversions says that bool and non-bool AltiVec vectors
+  // can be mixed, with the result being the non-bool type.  The non-bool
+  // operand must have integer element type.
+  if (AllowBoolConversions && LHSVecType && RHSVecType &&
+      LHSVecType->getNumElements() == RHSVecType->getNumElements() &&
+      (Context.getTypeSize(LHSVecType->getElementType()) ==
+       Context.getTypeSize(RHSVecType->getElementType()))) {
+    if (LHSVecType->getVectorKind() == VectorType::AltiVecVector &&
+        LHSVecType->getElementType()->isIntegerType() &&
+        RHSVecType->getVectorKind() == VectorType::AltiVecBool) {
+      RHS = ImpCastExprToType(RHS.get(), LHSType, CK_BitCast);
+      return LHSType;
+    }
+    if (!IsCompAssign &&
+        LHSVecType->getVectorKind() == VectorType::AltiVecBool &&
+        RHSVecType->getVectorKind() == VectorType::AltiVecVector &&
+        RHSVecType->getElementType()->isIntegerType()) {
+      LHS = ImpCastExprToType(LHS.get(), RHSType, CK_BitCast);
+      return RHSType;
+    }
+  }
+
   // If there's an ext-vector type and a scalar, try to convert the scalar to
   // the vector element type and splat.
   if (!RHSVecType && isa<ExtVectorType>(LHSVecType)) {
@@ -7391,7 +7426,9 @@
 
   if (LHS.get()->getType()->isVectorType() ||
       RHS.get()->getType()->isVectorType())
-    return CheckVectorOperands(LHS, RHS, Loc, IsCompAssign);
+    return CheckVectorOperands(LHS, RHS, Loc, IsCompAssign,
+                               /*AllowBothBool*/getLangOpts().AltiVec,
+                               /*AllowBoolConversions*/false);
 
   QualType compType = UsualArithmeticConversions(LHS, RHS, IsCompAssign);
   if (LHS.isInvalid() || RHS.isInvalid())
@@ -7420,7 +7457,9 @@
       RHS.get()->getType()->isVectorType()) {
     if (LHS.get()->getType()->hasIntegerRepresentation() && 
         RHS.get()->getType()->hasIntegerRepresentation())
-      return CheckVectorOperands(LHS, RHS, Loc, IsCompAssign);
+      return CheckVectorOperands(LHS, RHS, Loc, IsCompAssign,
+                                 /*AllowBothBool*/getLangOpts().AltiVec,
+                                 /*AllowBoolConversions*/false);
     return InvalidOperands(Loc, LHS, RHS);
   }
 
@@ -7706,7 +7745,10 @@
 
   if (LHS.get()->getType()->isVectorType() ||
       RHS.get()->getType()->isVectorType()) {
-    QualType compType = CheckVectorOperands(LHS, RHS, Loc, CompLHSTy);
+    QualType compType = CheckVectorOperands(
+        LHS, RHS, Loc, CompLHSTy,
+        /*AllowBothBool*/getLangOpts().AltiVec,
+        /*AllowBoolConversions*/getLangOpts().ZVector);
     if (CompLHSTy) *CompLHSTy = compType;
     return compType;
   }
@@ -7781,7 +7823,10 @@
 
   if (LHS.get()->getType()->isVectorType() ||
       RHS.get()->getType()->isVectorType()) {
-    QualType compType = CheckVectorOperands(LHS, RHS, Loc, CompLHSTy);
+    QualType compType = CheckVectorOperands(
+        LHS, RHS, Loc, CompLHSTy,
+        /*AllowBothBool*/getLangOpts().AltiVec,
+        /*AllowBoolConversions*/getLangOpts().ZVector);
     if (CompLHSTy) *CompLHSTy = compType;
     return compType;
   }
@@ -8023,7 +8068,21 @@
       RHS.get()->getType()->isVectorType()) {
     if (LangOpts.OpenCL)
       return checkOpenCLVectorShift(*this, LHS, RHS, Loc, IsCompAssign);
-    return CheckVectorOperands(LHS, RHS, Loc, IsCompAssign);
+    if (LangOpts.ZVector) {
+      // The shift operators for the z vector extensions work basically
+      // like OpenCL shifts, except that neither the LHS nor the RHS is
+      // allowed to be a "vector bool".
+      if (auto LHSVecType = LHS.get()->getType()->getAs<VectorType>())
+        if (LHSVecType->getVectorKind() == VectorType::AltiVecBool)
+          return InvalidOperands(Loc, LHS, RHS);
+      if (auto RHSVecType = RHS.get()->getType()->getAs<VectorType>())
+        if (RHSVecType->getVectorKind() == VectorType::AltiVecBool)
+          return InvalidOperands(Loc, LHS, RHS);
+      return checkOpenCLVectorShift(*this, LHS, RHS, Loc, IsCompAssign);
+    }
+    return CheckVectorOperands(LHS, RHS, Loc, IsCompAssign,
+                               /*AllowBothBool*/true,
+                               /*AllowBoolConversions*/false);
   }
 
   // Shifts don't perform usual arithmetic conversions, they just do integer
@@ -8797,7 +8856,9 @@
                                           bool IsRelational) {
   // Check to make sure we're operating on vectors of the same type and width,
   // Allowing one side to be a scalar of element type.
-  QualType vType = CheckVectorOperands(LHS, RHS, Loc, /*isCompAssign*/false);
+  QualType vType = CheckVectorOperands(LHS, RHS, Loc, /*isCompAssign*/false,
+                              /*AllowBothBool*/true,
+                              /*AllowBoolConversions*/getLangOpts().ZVector);
   if (vType.isNull())
     return vType;
 
@@ -8805,7 +8866,8 @@
 
   // If AltiVec, the comparison results in a numeric type, i.e.
   // bool for C++, int for C
-  if (vType->getAs<VectorType>()->getVectorKind() == VectorType::AltiVecVector)
+  if (getLangOpts().AltiVec &&
+      vType->getAs<VectorType>()->getVectorKind() == VectorType::AltiVecVector)
     return Context.getLogicalOperationType();
 
   // For non-floating point types, check for self-comparisons of the form
@@ -8839,7 +8901,9 @@
                                           SourceLocation Loc) {
   // Ensure that either both operands are of the same vector type, or
   // one operand is of a vector type and the other is of its element type.
-  QualType vType = CheckVectorOperands(LHS, RHS, Loc, false);
+  QualType vType = CheckVectorOperands(LHS, RHS, Loc, false,
+                                       /*AllowBothBool*/true,
+                                       /*AllowBoolConversions*/false);
   if (vType.isNull())
     return InvalidOperands(Loc, LHS, RHS);
   if (getLangOpts().OpenCL && getLangOpts().OpenCLVersion < 120 &&
@@ -8857,8 +8921,9 @@
       RHS.get()->getType()->isVectorType()) {
     if (LHS.get()->getType()->hasIntegerRepresentation() &&
         RHS.get()->getType()->hasIntegerRepresentation())
-      return CheckVectorOperands(LHS, RHS, Loc, IsCompAssign);
-    
+      return CheckVectorOperands(LHS, RHS, Loc, IsCompAssign,
+                        /*AllowBothBool*/true,
+                        /*AllowBoolConversions*/getLangOpts().ZVector);
     return InvalidOperands(Loc, LHS, RHS);
   }
 
@@ -9472,6 +9537,10 @@
                                           IsInc, IsPrefix);
   } else if (S.getLangOpts().AltiVec && ResType->isVectorType()) {
     // OK! ( C/C++ Language Extensions for CBEA(Version 2.6) 10.3 )
+  } else if (S.getLangOpts().ZVector && ResType->isVectorType() &&
+             (ResType->getAs<VectorType>()->getVectorKind() !=
+              VectorType::AltiVecBool)) {
+    // The z vector extensions allow ++ and -- for non-bool vectors.
   } else if(S.getLangOpts().OpenCL && ResType->isVectorType() &&
             ResType->getAs<VectorType>()->getElementType()->isIntegerType()) {
     // OpenCL V1.2 6.3 says dec/inc ops operate on integer vector types.
@@ -10552,8 +10621,13 @@
     resultType = Input.get()->getType();
     if (resultType->isDependentType())
       break;
-    if (resultType->isArithmeticType() || // C99 6.5.3.3p1
-        resultType->isVectorType()) 
+    if (resultType->isArithmeticType()) // C99 6.5.3.3p1
+      break;
+    else if (resultType->isVectorType() &&
+             // The z vector extensions don't allow + or - with bool vectors.
+             (!Context.getLangOpts().ZVector ||
+              resultType->getAs<VectorType>()->getVectorKind() !=
+              VectorType::AltiVecBool))
       break;
     else if (getLangOpts().CPlusPlus && // C++ [expr.unary.op]p6
              Opc == UO_Plus &&
diff --git a/clang/lib/Sema/SemaExprCXX.cpp b/clang/lib/Sema/SemaExprCXX.cpp
index 6608d7c..01966d5 100644
--- a/clang/lib/Sema/SemaExprCXX.cpp
+++ b/clang/lib/Sema/SemaExprCXX.cpp
@@ -4950,7 +4950,9 @@
 
   // Extension: conditional operator involving vector types.
   if (LTy->isVectorType() || RTy->isVectorType())
-    return CheckVectorOperands(LHS, RHS, QuestionLoc, /*isCompAssign*/false);
+    return CheckVectorOperands(LHS, RHS, QuestionLoc, /*isCompAssign*/false,
+                               /*AllowBothBool*/true,
+                               /*AllowBoolConversions*/false);
 
   //   -- The second and third operands have arithmetic or enumeration type;
   //      the usual arithmetic conversions are performed to bring them to a
diff --git a/clang/lib/Sema/SemaLookup.cpp b/clang/lib/Sema/SemaLookup.cpp
index 3fd1f21..2e7f891 100644
--- a/clang/lib/Sema/SemaLookup.cpp
+++ b/clang/lib/Sema/SemaLookup.cpp
@@ -4245,7 +4245,7 @@
   // Don't try to correct the identifier "vector" when in AltiVec mode.
   // TODO: Figure out why typo correction misbehaves in this case, fix it, and
   // remove this workaround.
-  if (getLangOpts().AltiVec && Typo->isStr("vector"))
+  if ((getLangOpts().AltiVec || getLangOpts().ZVector) && Typo->isStr("vector"))
     return nullptr;
 
   // Provide a stop gap for files that are just seriously broken.  Trying
diff --git a/clang/test/CodeGen/builtins-systemz-zvector-error.c b/clang/test/CodeGen/builtins-systemz-zvector-error.c
new file mode 100644
index 0000000..8d5380d
--- /dev/null
+++ b/clang/test/CodeGen/builtins-systemz-zvector-error.c
@@ -0,0 +1,576 @@
+// REQUIRES: systemz-registered-target
+// RUN: %clang_cc1 -target-cpu z13 -triple s390x-linux-gnu \
+// RUN: -fzvector -fno-lax-vector-conversions \
+// RUN: -Wall -Wno-unused -Werror -fsyntax-only -verify %s
+
+#include <vecintrin.h>
+
+volatile vector signed char vsc;
+volatile vector signed short vss;
+volatile vector signed int vsi;
+volatile vector signed long long vsl;
+volatile vector unsigned char vuc;
+volatile vector unsigned short vus;
+volatile vector unsigned int vui;
+volatile vector unsigned long long vul;
+volatile vector bool char vbc;
+volatile vector bool short vbs;
+volatile vector bool int vbi;
+volatile vector bool long long vbl;
+volatile vector double vd;
+
+volatile signed char sc;
+volatile signed short ss;
+volatile signed int si;
+volatile signed long long sl;
+volatile unsigned char uc;
+volatile unsigned short us;
+volatile unsigned int ui;
+volatile unsigned long long ul;
+volatile double d;
+
+const void * volatile cptr;
+const signed char * volatile cptrsc;
+const signed short * volatile cptrss;
+const signed int * volatile cptrsi;
+const signed long long * volatile cptrsl;
+const unsigned char * volatile cptruc;
+const unsigned short * volatile cptrus;
+const unsigned int * volatile cptrui;
+const unsigned long long * volatile cptrul;
+const float * volatile cptrf;
+const double * volatile cptrd;
+
+void * volatile ptr;
+signed char * volatile ptrsc;
+signed short * volatile ptrss;
+signed int * volatile ptrsi;
+signed long long * volatile ptrsl;
+unsigned char * volatile ptruc;
+unsigned short * volatile ptrus;
+unsigned int * volatile ptrui;
+unsigned long long * volatile ptrul;
+float * volatile ptrf;
+double * volatile ptrd;
+
+volatile unsigned int len;
+volatile int idx;
+int cc;
+
+void test_core(void) {
+  len = __lcbb(cptr, idx);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* {{must be a constant power of 2 from 64 to 4096}}
+  len = __lcbb(cptr, 200);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* {{must be a constant power of 2 from 64 to 4096}}
+  len = __lcbb(cptr, 32);    // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* {{must be a constant power of 2 from 64 to 4096}}
+  len = __lcbb(cptr, 8192);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* {{must be a constant power of 2 from 64 to 4096}}
+
+  vsl = vec_permi(vsl, vsl, idx); // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 3 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsl = vec_permi(vsl, vsl, -1);  // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 3 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsl = vec_permi(vsl, vsl, 4);   // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 3 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vul = vec_permi(vul, vul, idx); // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 2 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vul = vec_permi(vul, vul, -1);  // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 2 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vul = vec_permi(vul, vul, 4);   // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 2 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbl = vec_permi(vbl, vbl, idx); // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 2 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbl = vec_permi(vbl, vbl, -1);  // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 2 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbl = vec_permi(vbl, vbl, 4);   // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 2 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vd = vec_permi(vd, vd, idx);    // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 3 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vd = vec_permi(vd, vd, -1);     // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 3 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vd = vec_permi(vd, vd, 4);      // expected-error {{no matching function}}
+                                  // expected-note@vecintrin.h:* 3 {{candidate function not viable}}
+                                  // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+
+  vsi = vec_gather_element(vsi, vui, cptrsi, idx); // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsi = vec_gather_element(vsi, vui, cptrsi, -1);  // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsi = vec_gather_element(vsi, vui, cptrsi, 4);   // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vui = vec_gather_element(vui, vui, cptrui, idx); // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vui = vec_gather_element(vui, vui, cptrui, -1);  // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vui = vec_gather_element(vui, vui, cptrui, 4);   // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbi = vec_gather_element(vbi, vui, cptrui, idx); // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbi = vec_gather_element(vbi, vui, cptrui, -1);  // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbi = vec_gather_element(vbi, vui, cptrui, 4);   // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vsl = vec_gather_element(vsl, vul, cptrsl, idx); // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vsl = vec_gather_element(vsl, vul, cptrsl, -1);  // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vsl = vec_gather_element(vsl, vul, cptrsl, 2);   // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vul = vec_gather_element(vul, vul, cptrul, idx); // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vul = vec_gather_element(vul, vul, cptrul, -1);  // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vul = vec_gather_element(vul, vul, cptrul, 2);   // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vbl = vec_gather_element(vbl, vul, cptrul, idx); // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vbl = vec_gather_element(vbl, vul, cptrul, -1);  // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vbl = vec_gather_element(vbl, vul, cptrul, 2);   // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vd = vec_gather_element(vd, vul, cptrd, idx);    // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vd = vec_gather_element(vd, vul, cptrd, -1);     // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vd = vec_gather_element(vd, vul, cptrd, 2);      // expected-error {{no matching function}}
+                                                   // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                                   // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+
+  vec_scatter_element(vsi, vui, ptrsi, idx); // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vsi, vui, ptrsi, -1);  // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vsi, vui, ptrsi, 4);   // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vui, vui, ptrui, idx); // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vui, vui, ptrui, -1);  // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vui, vui, ptrui, 4);   // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vbi, vui, ptrui, idx); // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vbi, vui, ptrui, -1);  // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vbi, vui, ptrui, 4);   // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vec_scatter_element(vsl, vul, ptrsl, idx); // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vsl, vul, ptrsl, -1);  // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vsl, vul, ptrsl, 2);   // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vul, vul, ptrul, idx); // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vul, vul, ptrul, -1);  // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vul, vul, ptrul, 2);   // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vbl, vul, ptrul, idx); // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vbl, vul, ptrul, -1);  // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vbl, vul, ptrul, 2);   // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 5 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vd, vul, ptrd, idx);   // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vd, vul, ptrd, -1);    // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vec_scatter_element(vd, vul, ptrd, 2);     // expected-error {{no matching function}}
+                                             // expected-note@vecintrin.h:* 6 {{candidate function not viable}}
+                                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+
+  vsc = vec_load_bndry(cptrsc, idx);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vsc = vec_load_bndry(cptrsc, 200);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vsc = vec_load_bndry(cptrsc, 32);    // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vsc = vec_load_bndry(cptrsc, 8192);  // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vuc = vec_load_bndry(cptruc, idx);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vss = vec_load_bndry(cptrss, idx);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vus = vec_load_bndry(cptrus, idx);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vsi = vec_load_bndry(cptrsi, idx);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vui = vec_load_bndry(cptrui, idx);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vsl = vec_load_bndry(cptrsl, idx);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+  vul = vec_load_bndry(cptrul, idx);   // expected-error {{no matching function}}
+                                       // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                       // expected-note@vecintrin.h:* 1 {{must be a constant power of 2 from 64 to 4096}}
+
+  vuc = vec_genmask(idx);  // expected-error {{no matching function}}
+                           // expected-note@vecintrin.h:* {{must be a constant integer}}
+
+  vuc = vec_genmasks_8(0, idx);    // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vuc = vec_genmasks_8(idx, 0);    // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vuc = vec_genmasks_8(idx, idx);  // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vus = vec_genmasks_16(0, idx);   // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vus = vec_genmasks_16(idx, 0);   // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vus = vec_genmasks_16(idx, idx); // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vui = vec_genmasks_32(0, idx);   // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vui = vec_genmasks_32(idx, 0);   // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vui = vec_genmasks_32(idx, idx); // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vul = vec_genmasks_64(0, idx);   // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vul = vec_genmasks_64(idx, 0);   // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vul = vec_genmasks_64(idx, idx); // expected-error {{no matching function}}
+                                   // expected-note@vecintrin.h:* {{must be a constant integer}}
+
+  vsc = vec_splat(vsc, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vsc = vec_splat(vsc, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vsc = vec_splat(vsc, 16);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vuc = vec_splat(vuc, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 15}}
+  vuc = vec_splat(vuc, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 15}}
+  vuc = vec_splat(vuc, 16);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 15}}
+  vbc = vec_splat(vbc, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 15}}
+  vbc = vec_splat(vbc, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 15}}
+  vbc = vec_splat(vbc, 16);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 15}}
+  vss = vec_splat(vss, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 7}}
+  vss = vec_splat(vss, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 7}}
+  vss = vec_splat(vss, 8);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 7}}
+  vus = vec_splat(vus, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 7}}
+  vus = vec_splat(vus, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 7}}
+  vus = vec_splat(vus, 8);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 7}}
+  vbs = vec_splat(vbs, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 7}}
+  vbs = vec_splat(vbs, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 7}}
+  vbs = vec_splat(vbs, 8);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 7}}
+  vsi = vec_splat(vsi, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsi = vec_splat(vsi, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsi = vec_splat(vsi, 4);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vui = vec_splat(vui, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vui = vec_splat(vui, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vui = vec_splat(vui, 4);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbi = vec_splat(vbi, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbi = vec_splat(vbi, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vbi = vec_splat(vbi, 4);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 3}}
+  vsl = vec_splat(vsl, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vsl = vec_splat(vsl, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vsl = vec_splat(vsl, 2);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vul = vec_splat(vul, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vul = vec_splat(vul, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vul = vec_splat(vul, 2);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vbl = vec_splat(vbl, idx); // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vbl = vec_splat(vbl, -1);  // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vbl = vec_splat(vbl, 2);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 11 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 2 {{must be a constant integer from 0 to 1}}
+  vd = vec_splat(vd, idx);   // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vd = vec_splat(vd, -1);    // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+  vd = vec_splat(vd, 2);     // expected-error {{no matching function}}
+                             // expected-note@vecintrin.h:* 12 {{candidate function not viable}}
+                             // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 1}}
+
+  vsc = vec_splat_s8(idx);  // expected-error {{no matching function}}
+                            // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vuc = vec_splat_u8(idx);  // expected-error {{no matching function}}
+                            // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vss = vec_splat_s16(idx); // expected-error {{no matching function}}
+                            // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vus = vec_splat_u16(idx); // expected-error {{no matching function}}
+                            // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vsi = vec_splat_s32(idx); // expected-error {{no matching function}}
+                            // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vui = vec_splat_u32(idx); // expected-error {{no matching function}}
+                            // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vsl = vec_splat_s64(idx); // expected-error {{no matching function}}
+                            // expected-note@vecintrin.h:* {{must be a constant integer}}
+  vul = vec_splat_u64(idx); // expected-error {{no matching function}}
+                            // expected-note@vecintrin.h:* {{must be a constant integer}}
+}
+
+void test_integer(void) {
+  vsc = vec_rl_mask(vsc, vuc, idx); // expected-error {{no matching function}}
+                                    // expected-note@vecintrin.h:* 7 {{candidate function not viable}}
+                                    // expected-note@vecintrin.h:* 1 {{must be a constant integer}}
+  vuc = vec_rl_mask(vuc, vuc, idx); // expected-error {{no matching function}}
+                                    // expected-note@vecintrin.h:* 7 {{candidate function not viable}}
+                                    // expected-note@vecintrin.h:* 1 {{must be a constant integer}}
+  vss = vec_rl_mask(vss, vus, idx); // expected-error {{no matching function}}
+                                    // expected-note@vecintrin.h:* 7 {{candidate function not viable}}
+                                    // expected-note@vecintrin.h:* 1 {{must be a constant integer}}
+  vus = vec_rl_mask(vus, vus, idx); // expected-error {{no matching function}}
+                                    // expected-note@vecintrin.h:* 7 {{candidate function not viable}}
+                                    // expected-note@vecintrin.h:* 1 {{must be a constant integer}}
+  vsi = vec_rl_mask(vsi, vui, idx); // expected-error {{no matching function}}
+                                    // expected-note@vecintrin.h:* 7 {{candidate function not viable}}
+                                    // expected-note@vecintrin.h:* 1 {{must be a constant integer}}
+  vui = vec_rl_mask(vui, vui, idx); // expected-error {{no matching function}}
+                                    // expected-note@vecintrin.h:* 7 {{candidate function not viable}}
+                                    // expected-note@vecintrin.h:* 1 {{must be a constant integer}}
+  vsl = vec_rl_mask(vsl, vul, idx); // expected-error {{no matching function}}
+                                    // expected-note@vecintrin.h:* 7 {{candidate function not viable}}
+                                    // expected-note@vecintrin.h:* 1 {{must be a constant integer}}
+  vul = vec_rl_mask(vul, vul, idx); // expected-error {{no matching function}}
+                                    // expected-note@vecintrin.h:* 7 {{candidate function not viable}}
+                                    // expected-note@vecintrin.h:* 1 {{must be a constant integer}}
+
+  vsc = vec_sld(vsc, vsc, idx); // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vsc = vec_sld(vsc, vsc, -1);  // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vsc = vec_sld(vsc, vsc, 16);  // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vuc = vec_sld(vuc, vuc, idx); // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vuc = vec_sld(vuc, vuc, -1);  // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vuc = vec_sld(vuc, vuc, 16);  // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vss = vec_sld(vss, vss, idx); // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vus = vec_sld(vus, vus, idx); // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vsi = vec_sld(vsi, vsi, idx); // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vui = vec_sld(vui, vui, idx); // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vsl = vec_sld(vsl, vsl, idx); // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vul = vec_sld(vul, vul, idx); // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+  vd = vec_sld(vd, vd, idx);    // expected-error {{no matching function}}
+                                // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 15}}
+
+  vsc = vec_sldw(vsc, vsc, idx); // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsc = vec_sldw(vsc, vsc, -1);  // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsc = vec_sldw(vsc, vsc, 4);   // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vuc = vec_sldw(vuc, vuc, idx); // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vuc = vec_sldw(vuc, vuc, -1);  // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vuc = vec_sldw(vuc, vuc, 4);   // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vss = vec_sldw(vss, vss, idx); // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vus = vec_sldw(vus, vus, idx); // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsi = vec_sldw(vsi, vsi, idx); // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vui = vec_sldw(vui, vui, idx); // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vsl = vec_sldw(vsl, vsl, idx); // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vul = vec_sldw(vul, vul, idx); // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+  vd = vec_sldw(vd, vd, idx);    // expected-error {{no matching function}}
+                                 // expected-note@vecintrin.h:* 8 {{candidate function not viable}}
+                                 // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 3}}
+}
+
+void test_float(void) {
+  vd = vec_ctd(vsl, idx); // expected-error {{no matching function}}
+                          // expected-note@vecintrin.h:* 1 {{candidate function not viable}}
+                          // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vd = vec_ctd(vsl, -1);  // expected-error {{no matching function}}
+                          // expected-note@vecintrin.h:* 1 {{candidate function not viable}}
+                          // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vd = vec_ctd(vsl, 32);  // expected-error {{no matching function}}
+                          // expected-note@vecintrin.h:* 1 {{candidate function not viable}}
+                          // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vd = vec_ctd(vul, idx); // expected-error {{no matching function}}
+                          // expected-note@vecintrin.h:* 1 {{candidate function not viable}}
+                          // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vd = vec_ctd(vul, -1);  // expected-error {{no matching function}}
+                          // expected-note@vecintrin.h:* 1 {{candidate function not viable}}
+                          // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vd = vec_ctd(vul, 32);  // expected-error {{no matching function}}
+                          // expected-note@vecintrin.h:* 1 {{candidate function not viable}}
+                          // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+
+  vsl = vec_ctsl(vd, idx); // expected-error {{no matching function}}
+                           // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vsl = vec_ctsl(vd, -1);  // expected-error {{no matching function}}
+                           // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vsl = vec_ctsl(vd, 32);  // expected-error {{no matching function}}
+                           // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vul = vec_ctul(vd, idx); // expected-error {{no matching function}}
+                           // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vul = vec_ctul(vd, -1);  // expected-error {{no matching function}}
+                           // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+  vul = vec_ctul(vd, 32);  // expected-error {{no matching function}}
+                           // expected-note@vecintrin.h:* 1 {{must be a constant integer from 0 to 31}}
+
+  vbl = vec_fp_test_data_class(vd, idx, &cc);  // expected-error {{must be a constant integer}}
+  vbl = vec_fp_test_data_class(vd, -1, &cc);   // expected-error {{should be a value from 0 to 4095}}
+  vbl = vec_fp_test_data_class(vd, 4096, &cc); // expected-error {{should be a value from 0 to 4095}}
+}
diff --git a/clang/test/CodeGen/builtins-systemz-zvector.c b/clang/test/CodeGen/builtins-systemz-zvector.c
new file mode 100644
index 0000000..6d554af
--- /dev/null
+++ b/clang/test/CodeGen/builtins-systemz-zvector.c
@@ -0,0 +1,2967 @@
+// REQUIRES: systemz-registered-target
+// RUN: %clang_cc1 -target-cpu z13 -triple s390x-linux-gnu \
+// RUN: -O -fzvector -fno-lax-vector-conversions \
+// RUN: -Wall -Wno-unused -Werror -emit-llvm %s -o - | FileCheck %s
+
+#include <vecintrin.h>
+
+volatile vector signed char vsc;
+volatile vector signed short vss;
+volatile vector signed int vsi;
+volatile vector signed long long vsl;
+volatile vector unsigned char vuc;
+volatile vector unsigned short vus;
+volatile vector unsigned int vui;
+volatile vector unsigned long long vul;
+volatile vector bool char vbc;
+volatile vector bool short vbs;
+volatile vector bool int vbi;
+volatile vector bool long long vbl;
+volatile vector double vd;
+
+volatile signed char sc;
+volatile signed short ss;
+volatile signed int si;
+volatile signed long long sl;
+volatile unsigned char uc;
+volatile unsigned short us;
+volatile unsigned int ui;
+volatile unsigned long long ul;
+volatile double d;
+
+const void * volatile cptr;
+const signed char * volatile cptrsc;
+const signed short * volatile cptrss;
+const signed int * volatile cptrsi;
+const signed long long * volatile cptrsl;
+const unsigned char * volatile cptruc;
+const unsigned short * volatile cptrus;
+const unsigned int * volatile cptrui;
+const unsigned long long * volatile cptrul;
+const float * volatile cptrf;
+const double * volatile cptrd;
+
+void * volatile ptr;
+signed char * volatile ptrsc;
+signed short * volatile ptrss;
+signed int * volatile ptrsi;
+signed long long * volatile ptrsl;
+unsigned char * volatile ptruc;
+unsigned short * volatile ptrus;
+unsigned int * volatile ptrui;
+unsigned long long * volatile ptrul;
+float * volatile ptrf;
+double * volatile ptrd;
+
+volatile unsigned int len;
+volatile int idx;
+int cc;
+
+void test_core(void) {
+  len = __lcbb(cptr, 64);
+  // CHECK: call i32 @llvm.s390.lcbb(i8* %{{.*}}, i32 0)
+  len = __lcbb(cptr, 128);
+  // CHECK: call i32 @llvm.s390.lcbb(i8* %{{.*}}, i32 1)
+  len = __lcbb(cptr, 256);
+  // CHECK: call i32 @llvm.s390.lcbb(i8* %{{.*}}, i32 2)
+  len = __lcbb(cptr, 512);
+  // CHECK: call i32 @llvm.s390.lcbb(i8* %{{.*}}, i32 3)
+  len = __lcbb(cptr, 1024);
+  // CHECK: call i32 @llvm.s390.lcbb(i8* %{{.*}}, i32 4)
+  len = __lcbb(cptr, 2048);
+  // CHECK: call i32 @llvm.s390.lcbb(i8* %{{.*}}, i32 5)
+  len = __lcbb(cptr, 4096);
+  // CHECK: call i32 @llvm.s390.lcbb(i8* %{{.*}}, i32 6)
+
+  sc = vec_extract(vsc, idx);
+  // CHECK: extractelement <16 x i8> %{{.*}}, i32 %{{.*}}
+  uc = vec_extract(vuc, idx);
+  // CHECK: extractelement <16 x i8> %{{.*}}, i32 %{{.*}}
+  uc = vec_extract(vbc, idx);
+  // CHECK: extractelement <16 x i8> %{{.*}}, i32 %{{.*}}
+  ss = vec_extract(vss, idx);
+  // CHECK: extractelement <8 x i16> %{{.*}}, i32 %{{.*}}
+  us = vec_extract(vus, idx);
+  // CHECK: extractelement <8 x i16> %{{.*}}, i32 %{{.*}}
+  us = vec_extract(vbs, idx);
+  // CHECK: extractelement <8 x i16> %{{.*}}, i32 %{{.*}}
+  si = vec_extract(vsi, idx);
+  // CHECK: extractelement <4 x i32> %{{.*}}, i32 %{{.*}}
+  ui = vec_extract(vui, idx);
+  // CHECK: extractelement <4 x i32> %{{.*}}, i32 %{{.*}}
+  ui = vec_extract(vbi, idx);
+  // CHECK: extractelement <4 x i32> %{{.*}}, i32 %{{.*}}
+  sl = vec_extract(vsl, idx);
+  // CHECK: extractelement <2 x i64> %{{.*}}, i32 %{{.*}}
+  ul = vec_extract(vul, idx);
+  // CHECK: extractelement <2 x i64> %{{.*}}, i32 %{{.*}}
+  ul = vec_extract(vbl, idx);
+  // CHECK: extractelement <2 x i64> %{{.*}}, i32 %{{.*}}
+  d = vec_extract(vd, idx);
+  // CHECK: extractelement <2 x double> %{{.*}}, i32 %{{.*}}
+
+  vsc = vec_insert(sc, vsc, idx);
+  // CHECK: insertelement <16 x i8> %{{.*}}, i8 %{{.*}}, i32 %{{.*}}
+  vuc = vec_insert(uc, vuc, idx);
+  // CHECK: insertelement <16 x i8> %{{.*}}, i8 %{{.*}}, i32 %{{.*}}
+  vuc = vec_insert(uc, vbc, idx);
+  // CHECK: insertelement <16 x i8> %{{.*}}, i8 %{{.*}}, i32 %{{.*}}
+  vss = vec_insert(ss, vss, idx);
+  // CHECK: insertelement <8 x i16> %{{.*}}, i16 %{{.*}}, i32 %{{.*}}
+  vus = vec_insert(us, vus, idx);
+  // CHECK: insertelement <8 x i16> %{{.*}}, i16 %{{.*}}, i32 %{{.*}}
+  vus = vec_insert(us, vbs, idx);
+  // CHECK: insertelement <8 x i16> %{{.*}}, i16 %{{.*}}, i32 %{{.*}}
+  vsi = vec_insert(si, vsi, idx);
+  // CHECK: insertelement <4 x i32> %{{.*}}, i32 %{{.*}}, i32 %{{.*}}
+  vui = vec_insert(ui, vui, idx);
+  // CHECK: insertelement <4 x i32> %{{.*}}, i32 %{{.*}}, i32 %{{.*}}
+  vui = vec_insert(ui, vbi, idx);
+  // CHECK: insertelement <4 x i32> %{{.*}}, i32 %{{.*}}, i32 %{{.*}}
+  vsl = vec_insert(sl, vsl, idx);
+  // CHECK: insertelement <2 x i64> %{{.*}}, i64 %{{.*}}, i32 %{{.*}}
+  vul = vec_insert(ul, vul, idx);
+  // CHECK: insertelement <2 x i64> %{{.*}}, i64 %{{.*}}, i32 %{{.*}}
+  vul = vec_insert(ul, vbl, idx);
+  // CHECK: insertelement <2 x i64> %{{.*}}, i64 %{{.*}}, i32 %{{.*}}
+  vd = vec_insert(d, vd, idx);
+  // CHECK: insertelement <2 x double> %{{.*}}, double %{{.*}}, i32 %{{.*}}
+
+  vsc = vec_promote(sc, idx);
+  // CHECK: insertelement <16 x i8> undef, i8 %{{.*}}, i32 %{{.*}}
+  vuc = vec_promote(uc, idx);
+  // CHECK: insertelement <16 x i8> undef, i8 %{{.*}}, i32 %{{.*}}
+  vss = vec_promote(ss, idx);
+  // CHECK: insertelement <8 x i16> undef, i16 %{{.*}}, i32 %{{.*}}
+  vus = vec_promote(us, idx);
+  // CHECK: insertelement <8 x i16> undef, i16 %{{.*}}, i32 %{{.*}}
+  vsi = vec_promote(si, idx);
+  // CHECK: insertelement <4 x i32> undef, i32 %{{.*}}, i32 %{{.*}}
+  vui = vec_promote(ui, idx);
+  // CHECK: insertelement <4 x i32> undef, i32 %{{.*}}, i32 %{{.*}}
+  vsl = vec_promote(sl, idx);
+  // CHECK: insertelement <2 x i64> undef, i64 %{{.*}}, i32 %{{.*}}
+  vul = vec_promote(ul, idx);
+  // CHECK: insertelement <2 x i64> undef, i64 %{{.*}}, i32 %{{.*}}
+  vd = vec_promote(d, idx);
+  // CHECK: insertelement <2 x double> undef, double %{{.*}}, i32 %{{.*}}
+
+  vsc = vec_insert_and_zero(cptrsc);
+  // CHECK: insertelement <16 x i8> <i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 undef, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>, i8 %{{.*}}, i32 7
+  vuc = vec_insert_and_zero(cptruc);
+  // CHECK: insertelement <16 x i8> <i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 undef, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>, i8 %{{.*}}, i32 7
+  vss = vec_insert_and_zero(cptrss);
+  // CHECK: insertelement <8 x i16> <i16 0, i16 0, i16 0, i16 undef, i16 0, i16 0, i16 0, i16 0>, i16 %{{.*}}, i32 3
+  vus = vec_insert_and_zero(cptrus);
+  // CHECK: insertelement <8 x i16> <i16 0, i16 0, i16 0, i16 undef, i16 0, i16 0, i16 0, i16 0>, i16 %{{.*}}, i32 3
+  vsi = vec_insert_and_zero(cptrsi);
+  // CHECK: insertelement <4 x i32> <i32 0, i32 undef, i32 0, i32 0>, i32 %{{.*}}, i32 1
+  vui = vec_insert_and_zero(cptrui);
+  // CHECK: insertelement <4 x i32> <i32 0, i32 undef, i32 0, i32 0>, i32 %{{.*}}, i32 1
+  vsl = vec_insert_and_zero(cptrsl);
+  // CHECK: insertelement <2 x i64> <i64 undef, i64 0>, i64 %{{.*}}, i32 0
+  vul = vec_insert_and_zero(cptrul);
+  // CHECK: insertelement <2 x i64> <i64 undef, i64 0>, i64 %{{.*}}, i32 0
+  vd = vec_insert_and_zero(cptrd);
+  // CHECK: insertelement <2 x double> <double undef, double 0.000000e+00>, double %{{.*}}, i32 0
+
+  vsc = vec_perm(vsc, vsc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_perm(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_perm(vbc, vbc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_perm(vss, vss, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_perm(vus, vus, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_perm(vbs, vbs, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_perm(vsi, vsi, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_perm(vui, vui, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_perm(vbi, vbi, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_perm(vsl, vsl, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_perm(vul, vul, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_perm(vbl, vbl, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vd = vec_perm(vd, vd, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vperm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsl = vec_permi(vsl, vsl, 0);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 0)
+  vsl = vec_permi(vsl, vsl, 1);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 1)
+  vsl = vec_permi(vsl, vsl, 2);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 4)
+  vsl = vec_permi(vsl, vsl, 3);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 5)
+  vul = vec_permi(vul, vul, 0);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 0)
+  vul = vec_permi(vul, vul, 1);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 1)
+  vul = vec_permi(vul, vul, 2);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 4)
+  vul = vec_permi(vul, vul, 3);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 5)
+  vbl = vec_permi(vbl, vbl, 0);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 0)
+  vbl = vec_permi(vbl, vbl, 1);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 1)
+  vbl = vec_permi(vbl, vbl, 2);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 4)
+  vbl = vec_permi(vbl, vbl, 3);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 5)
+  vd = vec_permi(vd, vd, 0);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 0)
+  vd = vec_permi(vd, vd, 1);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 1)
+  vd = vec_permi(vd, vd, 2);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 4)
+  vd = vec_permi(vd, vd, 3);
+  // CHECK: call <2 x i64> @llvm.s390.vpdi(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 5)
+
+  vsc = vec_sel(vsc, vsc, vuc);
+  vsc = vec_sel(vsc, vsc, vbc);
+  vuc = vec_sel(vuc, vuc, vuc);
+  vuc = vec_sel(vuc, vuc, vbc);
+  vbc = vec_sel(vbc, vbc, vuc);
+  vbc = vec_sel(vbc, vbc, vbc);
+  vss = vec_sel(vss, vss, vus);
+  vss = vec_sel(vss, vss, vbs);
+  vus = vec_sel(vus, vus, vus);
+  vus = vec_sel(vus, vus, vbs);
+  vbs = vec_sel(vbs, vbs, vus);
+  vbs = vec_sel(vbs, vbs, vbs);
+  vsi = vec_sel(vsi, vsi, vui);
+  vsi = vec_sel(vsi, vsi, vbi);
+  vui = vec_sel(vui, vui, vui);
+  vui = vec_sel(vui, vui, vbi);
+  vbi = vec_sel(vbi, vbi, vui);
+  vbi = vec_sel(vbi, vbi, vbi);
+  vsl = vec_sel(vsl, vsl, vul);
+  vsl = vec_sel(vsl, vsl, vbl);
+  vul = vec_sel(vul, vul, vul);
+  vul = vec_sel(vul, vul, vbl);
+  vbl = vec_sel(vbl, vbl, vul);
+  vbl = vec_sel(vbl, vbl, vbl);
+  vd = vec_sel(vd, vd, vul);
+  vd = vec_sel(vd, vd, vbl);
+
+  vsi = vec_gather_element(vsi, vui, cptrsi, 0);
+  vsi = vec_gather_element(vsi, vui, cptrsi, 1);
+  vsi = vec_gather_element(vsi, vui, cptrsi, 2);
+  vsi = vec_gather_element(vsi, vui, cptrsi, 3);
+  vui = vec_gather_element(vui, vui, cptrui, 0);
+  vui = vec_gather_element(vui, vui, cptrui, 1);
+  vui = vec_gather_element(vui, vui, cptrui, 2);
+  vui = vec_gather_element(vui, vui, cptrui, 3);
+  vbi = vec_gather_element(vbi, vui, cptrui, 0);
+  vbi = vec_gather_element(vbi, vui, cptrui, 1);
+  vbi = vec_gather_element(vbi, vui, cptrui, 2);
+  vbi = vec_gather_element(vbi, vui, cptrui, 3);
+  vsl = vec_gather_element(vsl, vul, cptrsl, 0);
+  vsl = vec_gather_element(vsl, vul, cptrsl, 1);
+  vul = vec_gather_element(vul, vul, cptrul, 0);
+  vul = vec_gather_element(vul, vul, cptrul, 1);
+  vbl = vec_gather_element(vbl, vul, cptrul, 0);
+  vbl = vec_gather_element(vbl, vul, cptrul, 1);
+  vd = vec_gather_element(vd, vul, cptrd, 0);
+  vd = vec_gather_element(vd, vul, cptrd, 1);
+
+  vec_scatter_element(vsi, vui, ptrsi, 0);
+  vec_scatter_element(vsi, vui, ptrsi, 1);
+  vec_scatter_element(vsi, vui, ptrsi, 2);
+  vec_scatter_element(vsi, vui, ptrsi, 3);
+  vec_scatter_element(vui, vui, ptrui, 0);
+  vec_scatter_element(vui, vui, ptrui, 1);
+  vec_scatter_element(vui, vui, ptrui, 2);
+  vec_scatter_element(vui, vui, ptrui, 3);
+  vec_scatter_element(vbi, vui, ptrui, 0);
+  vec_scatter_element(vbi, vui, ptrui, 1);
+  vec_scatter_element(vbi, vui, ptrui, 2);
+  vec_scatter_element(vbi, vui, ptrui, 3);
+  vec_scatter_element(vsl, vul, ptrsl, 0);
+  vec_scatter_element(vsl, vul, ptrsl, 1);
+  vec_scatter_element(vul, vul, ptrul, 0);
+  vec_scatter_element(vul, vul, ptrul, 1);
+  vec_scatter_element(vbl, vul, ptrul, 0);
+  vec_scatter_element(vbl, vul, ptrul, 1);
+  vec_scatter_element(vd, vul, ptrd, 0);
+  vec_scatter_element(vd, vul, ptrd, 1);
+
+  vsc = vec_xld2(idx, cptrsc);
+  vuc = vec_xld2(idx, cptruc);
+  vss = vec_xld2(idx, cptrss);
+  vus = vec_xld2(idx, cptrus);
+  vsi = vec_xld2(idx, cptrsi);
+  vui = vec_xld2(idx, cptrui);
+  vsl = vec_xld2(idx, cptrsl);
+  vul = vec_xld2(idx, cptrul);
+  vd = vec_xld2(idx, cptrd);
+
+  vsc = vec_xlw4(idx, cptrsc);
+  vuc = vec_xlw4(idx, cptruc);
+  vss = vec_xlw4(idx, cptrss);
+  vus = vec_xlw4(idx, cptrus);
+  vsi = vec_xlw4(idx, cptrsi);
+  vui = vec_xlw4(idx, cptrui);
+
+  vec_xstd2(vsc, idx, ptrsc);
+  vec_xstd2(vuc, idx, ptruc);
+  vec_xstd2(vss, idx, ptrss);
+  vec_xstd2(vus, idx, ptrus);
+  vec_xstd2(vsi, idx, ptrsi);
+  vec_xstd2(vui, idx, ptrui);
+  vec_xstd2(vsl, idx, ptrsl);
+  vec_xstd2(vul, idx, ptrul);
+  vec_xstd2(vd, idx, ptrd);
+
+  vec_xstw4(vsc, idx, ptrsc);
+  vec_xstw4(vuc, idx, ptruc);
+  vec_xstw4(vss, idx, ptrss);
+  vec_xstw4(vus, idx, ptrus);
+  vec_xstw4(vsi, idx, ptrsi);
+  vec_xstw4(vui, idx, ptrui);
+
+  vsc = vec_load_bndry(cptrsc, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vuc = vec_load_bndry(cptruc, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vss = vec_load_bndry(cptrss, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vus = vec_load_bndry(cptrus, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vsi = vec_load_bndry(cptrsi, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vui = vec_load_bndry(cptrui, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vsl = vec_load_bndry(cptrsl, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vul = vec_load_bndry(cptrul, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vd = vec_load_bndry(cptrd, 64);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 0)
+  vsc = vec_load_bndry(cptrsc, 128);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 1)
+  vsc = vec_load_bndry(cptrsc, 256);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 2)
+  vsc = vec_load_bndry(cptrsc, 512);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 3)
+  vsc = vec_load_bndry(cptrsc, 1024);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 4)
+  vsc = vec_load_bndry(cptrsc, 2048);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 5)
+  vsc = vec_load_bndry(cptrsc, 4096);
+  // CHECK: call <16 x i8> @llvm.s390.vlbb(i8* %{{.*}}, i32 6)
+
+  vsc = vec_load_len(cptrsc, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+  vuc = vec_load_len(cptruc, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+  vss = vec_load_len(cptrss, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+  vus = vec_load_len(cptrus, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+  vsi = vec_load_len(cptrsi, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+  vui = vec_load_len(cptrui, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+  vsl = vec_load_len(cptrsl, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+  vul = vec_load_len(cptrul, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+  vd = vec_load_len(cptrd, idx);
+  // CHECK: call <16 x i8> @llvm.s390.vll(i32 %{{.*}}, i8* %{{.*}})
+
+  vec_store_len(vsc, ptrsc, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+  vec_store_len(vuc, ptruc, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+  vec_store_len(vss, ptrss, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+  vec_store_len(vus, ptrus, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+  vec_store_len(vsi, ptrsi, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+  vec_store_len(vui, ptrui, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+  vec_store_len(vsl, ptrsl, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+  vec_store_len(vul, ptrul, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+  vec_store_len(vd, ptrd, idx);
+  // CHECK: call void @llvm.s390.vstl(<16 x i8> %{{.*}}, i32 %{{.*}}, i8* %{{.*}})
+
+  vsl = vec_load_pair(sl, sl);
+  vul = vec_load_pair(ul, ul);
+
+  vuc = vec_genmask(0);
+  // CHECK: <16 x i8> zeroinitializer
+  vuc = vec_genmask(0x8000);
+  // CHECK: <16 x i8> <i8 -1, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0>
+  vuc = vec_genmask(0xffff);
+  // CHECK: <16 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+
+  vuc = vec_genmasks_8(0, 7);
+  // CHECK: <16 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+  vuc = vec_genmasks_8(1, 4);
+  // CHECK: <16 x i8> <i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120, i8 120>
+  vuc = vec_genmasks_8(6, 2);
+  // CHECK: <16 x i8> <i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29, i8 -29>
+  vus = vec_genmasks_16(0, 15);
+  // CHECK: <8 x i16> <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+  vus = vec_genmasks_16(2, 11);
+  // CHECK: <8 x i16> <i16 16368, i16 16368, i16 16368, i16 16368, i16 16368, i16 16368, i16 16368, i16 16368>
+  vus = vec_genmasks_16(9, 2);
+  // CHECK:  <8 x i16> <i16 -8065, i16 -8065, i16 -8065, i16 -8065, i16 -8065, i16 -8065, i16 -8065, i16 -8065>
+  vui = vec_genmasks_32(0, 31);
+  // CHECK: <4 x i32> <i32 -1, i32 -1, i32 -1, i32 -1>
+  vui = vec_genmasks_32(7, 20);
+  // CHECK: <4 x i32> <i32 33552384, i32 33552384, i32 33552384, i32 33552384>
+  vui = vec_genmasks_32(25, 4);
+  // CHECK: <4 x i32> <i32 -134217601, i32 -134217601, i32 -134217601, i32 -134217601>
+  vul = vec_genmasks_64(0, 63);
+  // CHECK: <2 x i64> <i64 -1, i64 -1>
+  vul = vec_genmasks_64(3, 40);
+  // CHECK: <2 x i64> <i64 2305843009205305344, i64 2305843009205305344>
+  vul = vec_genmasks_64(30, 11);
+  // CHECK: <2 x i64> <i64 -4503582447501313, i64 -4503582447501313>
+
+  vsc = vec_splat(vsc, 0);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <16 x i32> zeroinitializer
+  vsc = vec_splat(vsc, 15);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <16 x i32> <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
+  vuc = vec_splat(vuc, 0);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <16 x i32> zeroinitializer
+  vuc = vec_splat(vuc, 15);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <16 x i32> <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
+  vbc = vec_splat(vbc, 0);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <16 x i32> zeroinitializer
+  vbc = vec_splat(vbc, 15);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <16 x i32> <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
+  vss = vec_splat(vss, 0);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <8 x i32> zeroinitializer
+  vss = vec_splat(vss, 7);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+  vus = vec_splat(vus, 0);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <8 x i32> zeroinitializer
+  vus = vec_splat(vus, 7);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+  vbs = vec_splat(vbs, 0);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <8 x i32> zeroinitializer
+  vbs = vec_splat(vbs, 7);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
+  vsi = vec_splat(vsi, 0);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <4 x i32> zeroinitializer
+  vsi = vec_splat(vsi, 3);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
+  vui = vec_splat(vui, 0);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <4 x i32> zeroinitializer
+  vui = vec_splat(vui, 3);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
+  vbi = vec_splat(vbi, 0);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <4 x i32> zeroinitializer
+  vbi = vec_splat(vbi, 3);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
+  vsl = vec_splat(vsl, 0);
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <2 x i32> zeroinitializer
+  vsl = vec_splat(vsl, 1);
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
+  vul = vec_splat(vul, 0);
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <2 x i32> zeroinitializer
+  vul = vec_splat(vul, 1);
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
+  vbl = vec_splat(vbl, 0);
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <2 x i32> zeroinitializer
+  vbl = vec_splat(vbl, 1);
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <2 x i32> <i32 1, i32 1>
+  vd = vec_splat(vd, 0);
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> undef, <2 x i32> zeroinitializer
+  vd = vec_splat(vd, 1);
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> undef, <2 x i32> <i32 1, i32 1>
+
+  vsc = vec_splat_s8(-128);
+  // CHECK: <16 x i8> <i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128, i8 -128>
+  vsc = vec_splat_s8(127);
+  // CHECK: <16 x i8> <i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127, i8 127>
+  vuc = vec_splat_u8(1);
+  // CHECK: <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
+  vuc = vec_splat_u8(254);
+  // CHECK: <16 x i8> <i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2, i8 -2>
+  vss = vec_splat_s16(-32768);
+  // CHECK: <8 x i16> <i16 -32768, i16 -32768, i16 -32768, i16 -32768, i16 -32768, i16 -32768, i16 -32768, i16 -32768>
+  vss = vec_splat_s16(32767);
+  // CHECK: <8 x i16> <i16 32767, i16 32767, i16 32767, i16 32767, i16 32767, i16 32767, i16 32767, i16 32767>
+  vus = vec_splat_u16(1);
+  // CHECK: <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
+  vus = vec_splat_u16(65534);
+  // CHECK: <8 x i16> <i16 -2, i16 -2, i16 -2, i16 -2, i16 -2, i16 -2, i16 -2, i16 -2>
+  vsi = vec_splat_s32(-32768);
+  // CHECK: <4 x i32> <i32 -32768, i32 -32768, i32 -32768, i32 -32768>
+  vsi = vec_splat_s32(32767);
+  // CHECK: <4 x i32> <i32 32767, i32 32767, i32 32767, i32 32767>
+  vui = vec_splat_u32(-32768);
+  // CHECK: <4 x i32> <i32 -32768, i32 -32768, i32 -32768, i32 -32768>
+  vui = vec_splat_u32(32767);
+  // CHECK: <4 x i32> <i32 32767, i32 32767, i32 32767, i32 32767>
+  vsl = vec_splat_s64(-32768);
+  // CHECK: <2 x i64> <i64 -32768, i64 -32768>
+  vsl = vec_splat_s64(32767);
+  // CHECK: <2 x i64> <i64 32767, i64 32767>
+  vul = vec_splat_u64(-32768);
+  // CHECK: <2 x i64> <i64 -32768, i64 -32768>
+  vul = vec_splat_u64(32767);
+  // CHECK: <2 x i64> <i64 32767, i64 32767>
+
+  vsc = vec_splats(sc);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <16 x i32> zeroinitializer
+  vuc = vec_splats(uc);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> undef, <16 x i32> zeroinitializer
+  vss = vec_splats(ss);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <8 x i32> zeroinitializer
+  vus = vec_splats(us);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> undef, <8 x i32> zeroinitializer
+  vsi = vec_splats(si);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <4 x i32> zeroinitializer
+  vui = vec_splats(ui);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> undef, <4 x i32> zeroinitializer
+  vsl = vec_splats(sl);
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <2 x i32> zeroinitializer
+  vul = vec_splats(ul);
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> undef, <2 x i32> zeroinitializer
+  vd = vec_splats(d);
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> undef, <2 x i32> zeroinitializer
+
+  vsl = vec_extend_s64(vsc);
+  vsl = vec_extend_s64(vss);
+  vsl = vec_extend_s64(vsi);
+
+  vsc = vec_mergeh(vsc, vsc);
+  // shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
+  vuc = vec_mergeh(vuc, vuc);
+  // shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
+  vbc = vec_mergeh(vbc, vbc);
+  // shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 2, i32 18, i32 3, i32 19, i32 4, i32 20, i32 5, i32 21, i32 6, i32 22, i32 7, i32 23>
+  vss = vec_mergeh(vss, vss);
+  // shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
+  vus = vec_mergeh(vus, vus);
+  // shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
+  vbs = vec_mergeh(vbs, vbs);
+  // shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
+  vsi = vec_mergeh(vsi, vsi);
+  // shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  vui = vec_mergeh(vui, vui);
+  // shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  vbi = vec_mergeh(vbi, vbi);
+  // shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
+  vsl = vec_mergeh(vsl, vsl);
+  // shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i32> <i32 0, i32 2>
+  vul = vec_mergeh(vul, vul);
+  // shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i32> <i32 0, i32 2>
+  vbl = vec_mergeh(vbl, vbl);
+  // shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i32> <i32 0, i32 2>
+  vd = vec_mergeh(vd, vd);
+  // shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x i32> <i32 0, i32 2>
+
+  vsc = vec_mergel(vsc, vsc);
+  // shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
+  vuc = vec_mergel(vuc, vuc);
+  // shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
+  vbc = vec_mergel(vbc, vbc);
+  // shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %a, <16 x i8> %b, <16 x i32> <i32 8, i32 24, i32 9, i32 25, i32 10, i32 26, i32 11, i32 27, i32 12, i32 28, i32 13, i32 29, i32 14, i32 30, i32 15, i32 31>
+  vss = vec_mergel(vss, vss);
+  // shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
+  vus = vec_mergel(vus, vus);
+  // shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
+  vbs = vec_mergel(vbs, vbs);
+  // shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
+  vsi = vec_mergel(vsi, vsi);
+  // shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <i32 2, i32 6, i32 3, i32 7>
+  vui = vec_mergel(vui, vui);
+  // shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <i32 2, i32 6, i32 3, i32 7>
+  vbi = vec_mergel(vbi, vbi);
+  // shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <i32 2, i32 6, i32 3, i32 7>
+  vsl = vec_mergel(vsl, vsl);
+  // shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <i32 1, i32 3>
+  vul = vec_mergel(vul, vul);
+  // shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <i32 1, i32 3>
+  vbl = vec_mergel(vbl, vbl);
+  // shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <i32 1, i32 3>
+  vd = vec_mergel(vd, vd);
+  // shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <i32 1, i32 3>
+
+  vsc = vec_pack(vss, vss);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31>
+  vuc = vec_pack(vus, vus);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31>
+  vbc = vec_pack(vbs, vbs);
+  // CHECK: shufflevector <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15, i32 17, i32 19, i32 21, i32 23, i32 25, i32 27, i32 29, i32 31>
+  vss = vec_pack(vsi, vsi);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
+  vus = vec_pack(vui, vui);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
+  vbs = vec_pack(vbi, vbi);
+  // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
+  vsi = vec_pack(vsl, vsl);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
+  vui = vec_pack(vul, vul);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
+  vbi = vec_pack(vbl, vbl);
+  // CHECK: shufflevector <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
+
+  vsc = vec_packs(vss, vss);
+  // CHECK: call <16 x i8> @llvm.s390.vpksh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vuc = vec_packs(vus, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vpklsh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vss = vec_packs(vsi, vsi);
+  // CHECK: call <8 x i16> @llvm.s390.vpksf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vus = vec_packs(vui, vui);
+  // CHECK: call <8 x i16> @llvm.s390.vpklsf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vsi = vec_packs(vsl, vsl);
+  // CHECK: call <4 x i32> @llvm.s390.vpksg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  vui = vec_packs(vul, vul);
+  // CHECK: call <4 x i32> @llvm.s390.vpklsg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vsc = vec_packs_cc(vss, vss, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vpkshs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vuc = vec_packs_cc(vus, vus, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vpklshs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vss = vec_packs_cc(vsi, vsi, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vpksfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vus = vec_packs_cc(vui, vui, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vpklsfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vsi = vec_packs_cc(vsl, vsl, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vpksgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  vui = vec_packs_cc(vul, vul, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vpklsgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vuc = vec_packsu(vss, vss);
+  // CHECK: call <16 x i8> @llvm.s390.vpklsh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vuc = vec_packsu(vus, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vpklsh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_packsu(vsi, vsi);
+  // CHECK: call <8 x i16> @llvm.s390.vpklsf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vus = vec_packsu(vui, vui);
+  // CHECK: call <8 x i16> @llvm.s390.vpklsf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_packsu(vsl, vsl);
+  // CHECK: call <4 x i32> @llvm.s390.vpklsg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  vui = vec_packsu(vul, vul);
+  // CHECK: call <4 x i32> @llvm.s390.vpklsg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vuc = vec_packsu_cc(vus, vus, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vpklshs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_packsu_cc(vui, vui, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vpklsfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_packsu_cc(vul, vul, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vpklsgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vss = vec_unpackh(vsc);
+  // CHECK: call <8 x i16> @llvm.s390.vuphb(<16 x i8> %{{.*}})
+  vus = vec_unpackh(vuc);
+  // CHECK: call <8 x i16> @llvm.s390.vuplhb(<16 x i8> %{{.*}})
+  vbs = vec_unpackh(vbc);
+  // CHECK: call <8 x i16> @llvm.s390.vuphb(<16 x i8> %{{.*}})
+  vsi = vec_unpackh(vss);
+  // CHECK: call <4 x i32> @llvm.s390.vuphh(<8 x i16> %{{.*}})
+  vui = vec_unpackh(vus);
+  // CHECK: call <4 x i32> @llvm.s390.vuplhh(<8 x i16> %{{.*}})
+  vbi = vec_unpackh(vbs);
+  // CHECK: call <4 x i32> @llvm.s390.vuphh(<8 x i16> %{{.*}})
+  vsl = vec_unpackh(vsi);
+  // CHECK: call <2 x i64> @llvm.s390.vuphf(<4 x i32> %{{.*}})
+  vul = vec_unpackh(vui);
+  // CHECK: call <2 x i64> @llvm.s390.vuplhf(<4 x i32> %{{.*}})
+  vbl = vec_unpackh(vbi);
+  // CHECK: call <2 x i64> @llvm.s390.vuphf(<4 x i32> %{{.*}})
+
+  vss = vec_unpackl(vsc);
+  // CHECK: call <8 x i16> @llvm.s390.vuplb(<16 x i8> %{{.*}})
+  vus = vec_unpackl(vuc);
+  // CHECK: call <8 x i16> @llvm.s390.vupllb(<16 x i8> %{{.*}})
+  vbs = vec_unpackl(vbc);
+  // CHECK: call <8 x i16> @llvm.s390.vuplb(<16 x i8> %{{.*}})
+  vsi = vec_unpackl(vss);
+  // CHECK: call <4 x i32> @llvm.s390.vuplhw(<8 x i16> %{{.*}})
+  vui = vec_unpackl(vus);
+  // CHECK: call <4 x i32> @llvm.s390.vupllh(<8 x i16> %{{.*}})
+  vbi = vec_unpackl(vbs);
+  // CHECK: call <4 x i32> @llvm.s390.vuplhw(<8 x i16> %{{.*}})
+  vsl = vec_unpackl(vsi);
+  // CHECK: call <2 x i64> @llvm.s390.vuplf(<4 x i32> %{{.*}})
+  vul = vec_unpackl(vui);
+  // CHECK: call <2 x i64> @llvm.s390.vupllf(<4 x i32> %{{.*}})
+  vbl = vec_unpackl(vbi);
+  // CHECK: call <2 x i64> @llvm.s390.vuplf(<4 x i32> %{{.*}})
+}
+
+void test_compare(void) {
+  vbc = vec_cmpeq(vsc, vsc);
+  // CHECK: icmp eq <16 x i8> %{{.*}}, %{{.*}}
+  vbc = vec_cmpeq(vuc, vuc);
+  // CHECK: icmp eq <16 x i8> %{{.*}}, %{{.*}}
+  vbc = vec_cmpeq(vbc, vbc);
+  // CHECK: icmp eq <16 x i8> %{{.*}}, %{{.*}}
+  vbs = vec_cmpeq(vss, vss);
+  // CHECK: icmp eq <8 x i16> %{{.*}}, %{{.*}}
+  vbs = vec_cmpeq(vus, vus);
+  // CHECK: icmp eq <8 x i16> %{{.*}}, %{{.*}}
+  vbs = vec_cmpeq(vbs, vbs);
+  // CHECK: icmp eq <8 x i16> %{{.*}}, %{{.*}}
+  vbi = vec_cmpeq(vsi, vsi);
+  // CHECK: icmp eq <4 x i32> %{{.*}}, %{{.*}}
+  vbi = vec_cmpeq(vui, vui);
+  // CHECK: icmp eq <4 x i32> %{{.*}}, %{{.*}}
+  vbi = vec_cmpeq(vbi, vbi);
+  // CHECK: icmp eq <4 x i32> %{{.*}}, %{{.*}}
+  vbl = vec_cmpeq(vsl, vsl);
+  // CHECK: icmp eq <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmpeq(vul, vul);
+  // CHECK: icmp eq <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmpeq(vbl, vbl);
+  // CHECK: icmp eq <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmpeq(vd, vd);
+  // CHECK: fcmp oeq <2 x double> %{{.*}}, %{{.*}}
+
+  vbc = vec_cmpge(vsc, vsc);
+  // CHECK: icmp sge <16 x i8> %{{.*}}, %{{.*}}
+  vbc = vec_cmpge(vuc, vuc);
+  // CHECK: icmp uge <16 x i8> %{{.*}}, %{{.*}}
+  vbs = vec_cmpge(vss, vss);
+  // CHECK: icmp sge <8 x i16> %{{.*}}, %{{.*}}
+  vbs = vec_cmpge(vus, vus);
+  // CHECK: icmp uge <8 x i16> %{{.*}}, %{{.*}}
+  vbi = vec_cmpge(vsi, vsi);
+  // CHECK: icmp sge <4 x i32> %{{.*}}, %{{.*}}
+  vbi = vec_cmpge(vui, vui);
+  // CHECK: icmp uge <4 x i32> %{{.*}}, %{{.*}}
+  vbl = vec_cmpge(vsl, vsl);
+  // CHECK: icmp sge <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmpge(vul, vul);
+  // CHECK: icmp uge <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmpge(vd, vd);
+  // CHECK: fcmp oge <2 x double> %{{.*}}, %{{.*}}
+
+  vbc = vec_cmpgt(vsc, vsc);
+  // CHECK: icmp sgt <16 x i8> %{{.*}}, %{{.*}}
+  vbc = vec_cmpgt(vuc, vuc);
+  // CHECK: icmp ugt <16 x i8> %{{.*}}, %{{.*}}
+  vbs = vec_cmpgt(vss, vss);
+  // CHECK: icmp sgt <8 x i16> %{{.*}}, %{{.*}}
+  vbs = vec_cmpgt(vus, vus);
+  // CHECK: icmp ugt <8 x i16> %{{.*}}, %{{.*}}
+  vbi = vec_cmpgt(vsi, vsi);
+  // CHECK: icmp sgt <4 x i32> %{{.*}}, %{{.*}}
+  vbi = vec_cmpgt(vui, vui);
+  // CHECK: icmp ugt <4 x i32> %{{.*}}, %{{.*}}
+  vbl = vec_cmpgt(vsl, vsl);
+  // CHECK: icmp sgt <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmpgt(vul, vul);
+  // CHECK: icmp ugt <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmpgt(vd, vd);
+  // CHECK: fcmp ogt <2 x double> %{{.*}}, %{{.*}}
+
+  vbc = vec_cmple(vsc, vsc);
+  // CHECK: icmp sle <16 x i8> %{{.*}}, %{{.*}}
+  vbc = vec_cmple(vuc, vuc);
+  // CHECK: icmp ule <16 x i8> %{{.*}}, %{{.*}}
+  vbs = vec_cmple(vss, vss);
+  // CHECK: icmp sle <8 x i16> %{{.*}}, %{{.*}}
+  vbs = vec_cmple(vus, vus);
+  // CHECK: icmp ule <8 x i16> %{{.*}}, %{{.*}}
+  vbi = vec_cmple(vsi, vsi);
+  // CHECK: icmp sle <4 x i32> %{{.*}}, %{{.*}}
+  vbi = vec_cmple(vui, vui);
+  // CHECK: icmp ule <4 x i32> %{{.*}}, %{{.*}}
+  vbl = vec_cmple(vsl, vsl);
+  // CHECK: icmp sle <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmple(vul, vul);
+  // CHECK: icmp ule <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmple(vd, vd);
+  // CHECK: fcmp ole <2 x double> %{{.*}}, %{{.*}}
+
+  vbc = vec_cmplt(vsc, vsc);
+  // CHECK: icmp slt <16 x i8> %{{.*}}, %{{.*}}
+  vbc = vec_cmplt(vuc, vuc);
+  // CHECK: icmp ult <16 x i8> %{{.*}}, %{{.*}}
+  vbs = vec_cmplt(vss, vss);
+  // CHECK: icmp slt <8 x i16> %{{.*}}, %{{.*}}
+  vbs = vec_cmplt(vus, vus);
+  // CHECK: icmp ult <8 x i16> %{{.*}}, %{{.*}}
+  vbi = vec_cmplt(vsi, vsi);
+  // CHECK: icmp slt <4 x i32> %{{.*}}, %{{.*}}
+  vbi = vec_cmplt(vui, vui);
+  // CHECK: icmp ult <4 x i32> %{{.*}}, %{{.*}}
+  vbl = vec_cmplt(vsl, vsl);
+  // CHECK: icmp slt <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmplt(vul, vul);
+  // CHECK: icmp ult <2 x i64> %{{.*}}, %{{.*}}
+  vbl = vec_cmplt(vd, vd);
+  // CHECK: fcmp olt <2 x double> %{{.*}}, %{{.*}}
+
+  idx = vec_all_eq(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_eq(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_eq(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_eq(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_eq(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_eq(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_eq(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_eq(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_eq(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_eq(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_eq(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_eq(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_eq(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_eq(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_eq(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_eq(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_eq(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_eq(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_eq(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_eq(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_eq(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_eq(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_eq(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_eq(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_eq(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_eq(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_eq(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_eq(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_eq(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfcedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_all_ne(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ne(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ne(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ne(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ne(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ne(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ne(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ne(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ne(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ne(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ne(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ne(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ne(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ne(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ne(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ne(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ne(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ne(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ne(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ne(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ne(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ne(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ne(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ne(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ne(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ne(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ne(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ne(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ne(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfcedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_all_ge(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ge(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ge(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ge(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ge(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ge(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ge(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_ge(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ge(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ge(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ge(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ge(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ge(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ge(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_ge(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ge(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ge(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ge(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ge(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ge(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ge(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_ge(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ge(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ge(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ge(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ge(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ge(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ge(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_ge(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_all_gt(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_gt(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_gt(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_gt(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_gt(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_gt(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_gt(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_gt(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_gt(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_gt(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_gt(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_gt(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_gt(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_gt(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_gt(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_gt(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_gt(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_gt(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_gt(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_gt(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_gt(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_gt(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_gt(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_gt(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_gt(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_gt(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_gt(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_gt(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_gt(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchdbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_all_le(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_le(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_le(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_le(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_le(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_le(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_le(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_le(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_le(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_le(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_le(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_le(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_le(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_le(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_le(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_le(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_le(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_le(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_le(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_le(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_le(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_le(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_le(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_le(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_le(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_le(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_le(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_le(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_le(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_all_lt(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_lt(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_lt(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_lt(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_lt(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_lt(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_lt(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_all_lt(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_lt(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_lt(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_lt(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_lt(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_lt(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_lt(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_all_lt(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_lt(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_lt(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_lt(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_lt(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_lt(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_lt(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_all_lt(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_lt(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_lt(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_lt(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_lt(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_lt(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_lt(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_all_lt(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchdbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_all_nge(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+  idx = vec_all_ngt(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchdbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+  idx = vec_all_nle(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+  idx = vec_all_nlt(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchdbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_all_nan(vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vftcidb(<2 x double> %{{.*}}, i32 15)
+  idx = vec_all_numeric(vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vftcidb(<2 x double> %{{.*}}, i32 15)
+
+  idx = vec_any_eq(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_eq(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_eq(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_eq(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_eq(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_eq(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_eq(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_eq(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_eq(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_eq(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_eq(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_eq(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_eq(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_eq(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_eq(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_eq(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_eq(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_eq(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_eq(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_eq(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_eq(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_eq(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_eq(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_eq(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_eq(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_eq(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_eq(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_eq(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_eq(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfcedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_any_ne(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ne(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ne(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ne(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ne(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ne(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ne(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vceqbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ne(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ne(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ne(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ne(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ne(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ne(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ne(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vceqhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ne(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ne(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ne(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ne(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ne(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ne(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ne(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vceqfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ne(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ne(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ne(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ne(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ne(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ne(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ne(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vceqgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ne(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfcedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_any_ge(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ge(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ge(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ge(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ge(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ge(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ge(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_ge(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ge(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ge(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ge(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ge(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ge(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ge(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_ge(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ge(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ge(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ge(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ge(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ge(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ge(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_ge(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ge(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ge(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ge(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ge(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ge(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ge(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_ge(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_any_gt(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_gt(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_gt(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_gt(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_gt(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_gt(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_gt(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_gt(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_gt(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_gt(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_gt(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_gt(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_gt(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_gt(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_gt(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_gt(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_gt(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_gt(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_gt(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_gt(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_gt(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_gt(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_gt(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_gt(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_gt(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_gt(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_gt(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_gt(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_gt(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchdbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_any_le(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_le(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_le(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_le(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_le(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_le(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_le(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_le(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_le(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_le(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_le(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_le(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_le(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_le(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_le(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_le(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_le(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_le(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_le(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_le(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_le(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_le(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_le(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_le(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_le(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_le(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_le(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_le(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_le(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_any_lt(vsc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_lt(vsc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_lt(vbc, vsc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_lt(vuc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_lt(vuc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_lt(vbc, vuc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_lt(vbc, vbc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vchlbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_any_lt(vss, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_lt(vss, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_lt(vbs, vss);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_lt(vus, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_lt(vus, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_lt(vbs, vus);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_lt(vbs, vbs);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vchlhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  idx = vec_any_lt(vsi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_lt(vsi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_lt(vbi, vsi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_lt(vui, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_lt(vui, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_lt(vbi, vui);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_lt(vbi, vbi);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vchlfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  idx = vec_any_lt(vsl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_lt(vsl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_lt(vbl, vsl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_lt(vul, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_lt(vul, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_lt(vbl, vul);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_lt(vbl, vbl);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vchlgs(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  idx = vec_any_lt(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchdbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_any_nge(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+  idx = vec_any_ngt(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchdbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+  idx = vec_any_nle(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchedbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+  idx = vec_any_nlt(vd, vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vfchdbs(<2 x double> %{{.*}}, <2 x double> %{{.*}})
+
+  idx = vec_any_nan(vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vftcidb(<2 x double> %{{.*}}, i32 15)
+  idx = vec_any_numeric(vd);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vftcidb(<2 x double> %{{.*}}, i32 15)
+}
+
+void test_integer(void) {
+  vsc = vec_andc(vsc, vsc);
+  vsc = vec_andc(vsc, vbc);
+  vsc = vec_andc(vbc, vsc);
+  vuc = vec_andc(vuc, vuc);
+  vuc = vec_andc(vuc, vbc);
+  vuc = vec_andc(vbc, vuc);
+  vbc = vec_andc(vbc, vbc);
+  vss = vec_andc(vss, vss);
+  vss = vec_andc(vss, vbs);
+  vss = vec_andc(vbs, vss);
+  vus = vec_andc(vus, vus);
+  vus = vec_andc(vus, vbs);
+  vus = vec_andc(vbs, vus);
+  vbs = vec_andc(vbs, vbs);
+  vsi = vec_andc(vsi, vsi);
+  vsi = vec_andc(vsi, vbi);
+  vsi = vec_andc(vbi, vsi);
+  vui = vec_andc(vui, vui);
+  vui = vec_andc(vui, vbi);
+  vui = vec_andc(vbi, vui);
+  vbi = vec_andc(vbi, vbi);
+  vsl = vec_andc(vsl, vsl);
+  vsl = vec_andc(vsl, vbl);
+  vsl = vec_andc(vbl, vsl);
+  vul = vec_andc(vul, vul);
+  vul = vec_andc(vul, vbl);
+  vul = vec_andc(vbl, vul);
+  vbl = vec_andc(vbl, vbl);
+  vd = vec_andc(vd, vd);
+  vd = vec_andc(vd, vbl);
+  vd = vec_andc(vbl, vd);
+
+  vsc = vec_nor(vsc, vsc);
+  vsc = vec_nor(vsc, vbc);
+  vsc = vec_nor(vbc, vsc);
+  vuc = vec_nor(vuc, vuc);
+  vuc = vec_nor(vuc, vbc);
+  vuc = vec_nor(vbc, vuc);
+  vbc = vec_nor(vbc, vbc);
+  vss = vec_nor(vss, vss);
+  vss = vec_nor(vss, vbs);
+  vss = vec_nor(vbs, vss);
+  vus = vec_nor(vus, vus);
+  vus = vec_nor(vus, vbs);
+  vus = vec_nor(vbs, vus);
+  vbs = vec_nor(vbs, vbs);
+  vsi = vec_nor(vsi, vsi);
+  vsi = vec_nor(vsi, vbi);
+  vsi = vec_nor(vbi, vsi);
+  vui = vec_nor(vui, vui);
+  vui = vec_nor(vui, vbi);
+  vui = vec_nor(vbi, vui);
+  vbi = vec_nor(vbi, vbi);
+  vsl = vec_nor(vsl, vsl);
+  vsl = vec_nor(vsl, vbl);
+  vsl = vec_nor(vbl, vsl);
+  vul = vec_nor(vul, vul);
+  vul = vec_nor(vul, vbl);
+  vul = vec_nor(vbl, vul);
+  vbl = vec_nor(vbl, vbl);
+  vd = vec_nor(vd, vd);
+  vd = vec_nor(vd, vbl);
+  vd = vec_nor(vbl, vd);
+
+  vuc = vec_cntlz(vsc);
+  // CHECK: call <16 x i8> @llvm.ctlz.v16i8(<16 x i8> %{{.*}}, i1 false)
+  vuc = vec_cntlz(vuc);
+  // CHECK: call <16 x i8> @llvm.ctlz.v16i8(<16 x i8> %{{.*}}, i1 false)
+  vus = vec_cntlz(vss);
+  // CHECK: call <8 x i16> @llvm.ctlz.v8i16(<8 x i16> %{{.*}}, i1 false)
+  vus = vec_cntlz(vus);
+  // CHECK: call <8 x i16> @llvm.ctlz.v8i16(<8 x i16> %{{.*}}, i1 false)
+  vui = vec_cntlz(vsi);
+  // CHECK: call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %{{.*}}, i1 false)
+  vui = vec_cntlz(vui);
+  // CHECK: call <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %{{.*}}, i1 false)
+  vul = vec_cntlz(vsl);
+  // CHECK: call <2 x i64> @llvm.ctlz.v2i64(<2 x i64> %{{.*}}, i1 false)
+  vul = vec_cntlz(vul);
+  // CHECK: call <2 x i64> @llvm.ctlz.v2i64(<2 x i64> %{{.*}}, i1 false)
+
+  vuc = vec_cnttz(vsc);
+  // CHECK: call <16 x i8> @llvm.cttz.v16i8(<16 x i8> %{{.*}}, i1 false)
+  vuc = vec_cnttz(vuc);
+  // CHECK: call <16 x i8> @llvm.cttz.v16i8(<16 x i8> %{{.*}}, i1 false)
+  vus = vec_cnttz(vss);
+  // CHECK: call <8 x i16> @llvm.cttz.v8i16(<8 x i16> %{{.*}}, i1 false)
+  vus = vec_cnttz(vus);
+  // CHECK: call <8 x i16> @llvm.cttz.v8i16(<8 x i16> %{{.*}}, i1 false)
+  vui = vec_cnttz(vsi);
+  // CHECK: call <4 x i32> @llvm.cttz.v4i32(<4 x i32> %{{.*}}, i1 false)
+  vui = vec_cnttz(vui);
+  // CHECK: call <4 x i32> @llvm.cttz.v4i32(<4 x i32> %{{.*}}, i1 false)
+  vul = vec_cnttz(vsl);
+  // CHECK: call <2 x i64> @llvm.cttz.v2i64(<2 x i64> %{{.*}}, i1 false)
+  vul = vec_cnttz(vul);
+  // CHECK: call <2 x i64> @llvm.cttz.v2i64(<2 x i64> %{{.*}}, i1 false)
+
+  vuc = vec_popcnt(vsc);
+  // CHECK: call <16 x i8> @llvm.ctpop.v16i8(<16 x i8> %{{.*}})
+  vuc = vec_popcnt(vuc);
+  // CHECK: call <16 x i8> @llvm.ctpop.v16i8(<16 x i8> %{{.*}})
+  vus = vec_popcnt(vss);
+  // CHECK: call <8 x i16> @llvm.ctpop.v8i16(<8 x i16> %{{.*}})
+  vus = vec_popcnt(vus);
+  // CHECK: call <8 x i16> @llvm.ctpop.v8i16(<8 x i16> %{{.*}})
+  vui = vec_popcnt(vsi);
+  // CHECK: call <4 x i32> @llvm.ctpop.v4i32(<4 x i32> %{{.*}})
+  vui = vec_popcnt(vui);
+  // CHECK: call <4 x i32> @llvm.ctpop.v4i32(<4 x i32> %{{.*}})
+  vul = vec_popcnt(vsl);
+  // CHECK: call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> %{{.*}})
+  vul = vec_popcnt(vul);
+  // CHECK: call <2 x i64> @llvm.ctpop.v2i64(<2 x i64> %{{.*}})
+
+  vsc = vec_rl(vsc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.verllvb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_rl(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.verllvb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_rl(vss, vus);
+  // CHECK: call <8 x i16> @llvm.s390.verllvh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_rl(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.verllvh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_rl(vsi, vui);
+  // CHECK: call <4 x i32> @llvm.s390.verllvf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_rl(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.verllvf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vsl = vec_rl(vsl, vul);
+  // CHECK: call <2 x i64> @llvm.s390.verllvg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  vul = vec_rl(vul, vul);
+  // CHECK: call <2 x i64> @llvm.s390.verllvg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vsc = vec_rli(vsc, ul);
+  // CHECK: call <16 x i8> @llvm.s390.verllb(<16 x i8> %{{.*}}, i32 %{{.*}})
+  vuc = vec_rli(vuc, ul);
+  // CHECK: call <16 x i8> @llvm.s390.verllb(<16 x i8> %{{.*}}, i32 %{{.*}})
+  vss = vec_rli(vss, ul);
+  // CHECK: call <8 x i16> @llvm.s390.verllh(<8 x i16> %{{.*}}, i32 %{{.*}})
+  vus = vec_rli(vus, ul);
+  // CHECK: call <8 x i16> @llvm.s390.verllh(<8 x i16> %{{.*}}, i32 %{{.*}})
+  vsi = vec_rli(vsi, ul);
+  // CHECK: call <4 x i32> @llvm.s390.verllf(<4 x i32> %{{.*}}, i32 %{{.*}})
+  vui = vec_rli(vui, ul);
+  // CHECK: call <4 x i32> @llvm.s390.verllf(<4 x i32> %{{.*}}, i32 %{{.*}})
+  vsl = vec_rli(vsl, ul);
+  // CHECK: call <2 x i64> @llvm.s390.verllg(<2 x i64> %{{.*}}, i32 %{{.*}})
+  vul = vec_rli(vul, ul);
+  // CHECK: call <2 x i64> @llvm.s390.verllg(<2 x i64> %{{.*}}, i32 %{{.*}})
+
+  vsc = vec_rl_mask(vsc, vuc, 0);
+  // CHECK: call <16 x i8> @llvm.s390.verimb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vsc = vec_rl_mask(vsc, vuc, 255);
+  // CHECK: call <16 x i8> @llvm.s390.verimb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 255)
+  vuc = vec_rl_mask(vuc, vuc, 0);
+  // CHECK: call <16 x i8> @llvm.s390.verimb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_rl_mask(vuc, vuc, 255);
+  // CHECK: call <16 x i8> @llvm.s390.verimb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 255)
+  vss = vec_rl_mask(vss, vus, 0);
+  // CHECK: call <8 x i16> @llvm.s390.verimh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vss = vec_rl_mask(vss, vus, 255);
+  // CHECK: call <8 x i16> @llvm.s390.verimh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 255)
+  vus = vec_rl_mask(vus, vus, 0);
+  // CHECK: call <8 x i16> @llvm.s390.verimh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_rl_mask(vus, vus, 255);
+  // CHECK: call <8 x i16> @llvm.s390.verimh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 255)
+  vsi = vec_rl_mask(vsi, vui, 0);
+  // CHECK: call <4 x i32> @llvm.s390.verimf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vsi = vec_rl_mask(vsi, vui, 255);
+  // CHECK: call <4 x i32> @llvm.s390.verimf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 255)
+  vui = vec_rl_mask(vui, vui, 0);
+  // CHECK: call <4 x i32> @llvm.s390.verimf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_rl_mask(vui, vui, 255);
+  // CHECK: call <4 x i32> @llvm.s390.verimf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 255)
+  vsl = vec_rl_mask(vsl, vul, 0);
+  // CHECK: call <2 x i64> @llvm.s390.verimg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 0)
+  vsl = vec_rl_mask(vsl, vul, 255);
+  // CHECK: call <2 x i64> @llvm.s390.verimg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 255)
+  vul = vec_rl_mask(vul, vul, 0);
+  // CHECK: call <2 x i64> @llvm.s390.verimg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 0)
+  vul = vec_rl_mask(vul, vul, 255);
+  // CHECK: call <2 x i64> @llvm.s390.verimg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, i32 255)
+
+  vsc = vec_sll(vsc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_sll(vsc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_sll(vsc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_sll(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_sll(vuc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_sll(vuc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_sll(vbc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_sll(vbc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_sll(vbc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_sll(vss, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_sll(vss, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_sll(vss, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_sll(vus, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_sll(vus, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_sll(vus, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_sll(vbs, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_sll(vbs, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_sll(vbs, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_sll(vsi, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_sll(vsi, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_sll(vsi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_sll(vui, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_sll(vui, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_sll(vui, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_sll(vbi, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_sll(vbi, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_sll(vbi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_sll(vsl, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_sll(vsl, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_sll(vsl, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_sll(vul, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_sll(vul, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_sll(vul, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_sll(vbl, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_sll(vbl, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_sll(vbl, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsc = vec_slb(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_slb(vsc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_slb(vuc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_slb(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_slb(vss, vss);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_slb(vss, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_slb(vus, vss);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_slb(vus, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_slb(vsi, vsi);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_slb(vsi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_slb(vui, vsi);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_slb(vui, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_slb(vsl, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_slb(vsl, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_slb(vul, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_slb(vul, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vd = vec_slb(vd, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vd = vec_slb(vd, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vslb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsc = vec_sld(vsc, vsc, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vsc = vec_sld(vsc, vsc, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+  vuc = vec_sld(vuc, vuc, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_sld(vuc, vuc, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+  vss = vec_sld(vss, vss, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vss = vec_sld(vss, vss, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+  vus = vec_sld(vus, vus, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vus = vec_sld(vus, vus, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+  vsi = vec_sld(vsi, vsi, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vsi = vec_sld(vsi, vsi, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+  vui = vec_sld(vui, vui, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vui = vec_sld(vui, vui, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+  vsl = vec_sld(vsl, vsl, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vsl = vec_sld(vsl, vsl, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+  vul = vec_sld(vul, vul, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vul = vec_sld(vul, vul, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+  vd = vec_sld(vd, vd, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vd = vec_sld(vd, vd, 15);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 15)
+
+  vsc = vec_sldw(vsc, vsc, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vsc = vec_sldw(vsc, vsc, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vuc = vec_sldw(vuc, vuc, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_sldw(vuc, vuc, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vss = vec_sldw(vss, vss, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vss = vec_sldw(vss, vss, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vus = vec_sldw(vus, vus, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vus = vec_sldw(vus, vus, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vsi = vec_sldw(vsi, vsi, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vsi = vec_sldw(vsi, vsi, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vui = vec_sldw(vui, vui, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vui = vec_sldw(vui, vui, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vsl = vec_sldw(vsl, vsl, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vsl = vec_sldw(vsl, vsl, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vul = vec_sldw(vul, vul, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vul = vec_sldw(vul, vul, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vd = vec_sldw(vd, vd, 0);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vd = vec_sldw(vd, vd, 3);
+  // CHECK: call <16 x i8> @llvm.s390.vsldb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+
+  vsc = vec_sral(vsc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_sral(vsc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_sral(vsc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_sral(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_sral(vuc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_sral(vuc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_sral(vbc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_sral(vbc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_sral(vbc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_sral(vss, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_sral(vss, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_sral(vss, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_sral(vus, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_sral(vus, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_sral(vus, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_sral(vbs, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_sral(vbs, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_sral(vbs, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_sral(vsi, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_sral(vsi, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_sral(vsi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_sral(vui, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_sral(vui, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_sral(vui, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_sral(vbi, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_sral(vbi, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_sral(vbi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_sral(vsl, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_sral(vsl, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_sral(vsl, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_sral(vul, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_sral(vul, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_sral(vul, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_sral(vbl, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_sral(vbl, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_sral(vbl, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsra(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsc = vec_srab(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_srab(vsc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_srab(vuc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_srab(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_srab(vss, vss);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_srab(vss, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_srab(vus, vss);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_srab(vus, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_srab(vsi, vsi);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_srab(vsi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_srab(vui, vsi);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_srab(vui, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_srab(vsl, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_srab(vsl, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_srab(vul, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_srab(vul, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vd = vec_srab(vd, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vd = vec_srab(vd, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vsrab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsc = vec_srl(vsc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_srl(vsc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_srl(vsc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_srl(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_srl(vuc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_srl(vuc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_srl(vbc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_srl(vbc, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbc = vec_srl(vbc, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_srl(vss, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_srl(vss, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_srl(vss, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_srl(vus, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_srl(vus, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_srl(vus, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_srl(vbs, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_srl(vbs, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbs = vec_srl(vbs, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_srl(vsi, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_srl(vsi, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_srl(vsi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_srl(vui, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_srl(vui, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_srl(vui, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_srl(vbi, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_srl(vbi, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbi = vec_srl(vbi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_srl(vsl, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_srl(vsl, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_srl(vsl, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_srl(vul, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_srl(vul, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_srl(vul, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_srl(vbl, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_srl(vbl, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vbl = vec_srl(vbl, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrl(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsc = vec_srb(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsc = vec_srb(vsc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_srb(vuc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_srb(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_srb(vss, vss);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_srb(vss, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_srb(vus, vss);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_srb(vus, vus);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_srb(vsi, vsi);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_srb(vsi, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_srb(vui, vsi);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_srb(vui, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_srb(vsl, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsl = vec_srb(vsl, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_srb(vul, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vul = vec_srb(vul, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vd = vec_srb(vd, vsl);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vd = vec_srb(vd, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vsrlb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsc = vec_abs(vsc);
+  vss = vec_abs(vss);
+  vsi = vec_abs(vsi);
+  vsl = vec_abs(vsl);
+
+  vsc = vec_max(vsc, vsc);
+  vsc = vec_max(vsc, vbc);
+  vsc = vec_max(vbc, vsc);
+  vuc = vec_max(vuc, vuc);
+  vuc = vec_max(vuc, vbc);
+  vuc = vec_max(vbc, vuc);
+  vss = vec_max(vss, vss);
+  vss = vec_max(vss, vbs);
+  vss = vec_max(vbs, vss);
+  vus = vec_max(vus, vus);
+  vus = vec_max(vus, vbs);
+  vus = vec_max(vbs, vus);
+  vsi = vec_max(vsi, vsi);
+  vsi = vec_max(vsi, vbi);
+  vsi = vec_max(vbi, vsi);
+  vui = vec_max(vui, vui);
+  vui = vec_max(vui, vbi);
+  vui = vec_max(vbi, vui);
+  vsl = vec_max(vsl, vsl);
+  vsl = vec_max(vsl, vbl);
+  vsl = vec_max(vbl, vsl);
+  vul = vec_max(vul, vul);
+  vul = vec_max(vul, vbl);
+  vul = vec_max(vbl, vul);
+  vd = vec_max(vd, vd);
+
+  vsc = vec_min(vsc, vsc);
+  vsc = vec_min(vsc, vbc);
+  vsc = vec_min(vbc, vsc);
+  vuc = vec_min(vuc, vuc);
+  vuc = vec_min(vuc, vbc);
+  vuc = vec_min(vbc, vuc);
+  vss = vec_min(vss, vss);
+  vss = vec_min(vss, vbs);
+  vss = vec_min(vbs, vss);
+  vus = vec_min(vus, vus);
+  vus = vec_min(vus, vbs);
+  vus = vec_min(vbs, vus);
+  vsi = vec_min(vsi, vsi);
+  vsi = vec_min(vsi, vbi);
+  vsi = vec_min(vbi, vsi);
+  vui = vec_min(vui, vui);
+  vui = vec_min(vui, vbi);
+  vui = vec_min(vbi, vui);
+  vsl = vec_min(vsl, vsl);
+  vsl = vec_min(vsl, vbl);
+  vsl = vec_min(vbl, vsl);
+  vul = vec_min(vul, vul);
+  vul = vec_min(vul, vbl);
+  vul = vec_min(vbl, vul);
+  vd = vec_min(vd, vd);
+
+  vuc = vec_addc(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vaccb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_addc(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vacch(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vui = vec_addc(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vaccf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vul = vec_addc(vul, vul);
+  // CHECK: call <2 x i64> @llvm.s390.vaccg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vuc = vec_add_u128(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vaq(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_addc_u128(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vaccq(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_adde_u128(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vacq(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_addec_u128(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vacccq(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsc = vec_avg(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vavgb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_avg(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vavglb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_avg(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vavgh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_avg(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vavglh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_avg(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vavgf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_avg(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vavglf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vsl = vec_avg(vsl, vsl);
+  // CHECK: call <2 x i64> @llvm.s390.vavgg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+  vul = vec_avg(vul, vul);
+  // CHECK: call <2 x i64> @llvm.s390.vavglg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vui = vec_checksum(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vcksm(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vus = vec_gfmsum(vuc, vuc);
+  // CHECK: call <8 x i16> @llvm.s390.vgfmb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_gfmsum(vus, vus);
+  // CHECK: call <4 x i32> @llvm.s390.vgfmh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vul = vec_gfmsum(vui, vui);
+  // CHECK: call <2 x i64> @llvm.s390.vgfmf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vuc = vec_gfmsum_128(vul, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vgfmg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vus = vec_gfmsum_accum(vuc, vuc, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vgfmab(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <8 x i16> %{{.*}})
+  vui = vec_gfmsum_accum(vus, vus, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vgfmah(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <4 x i32> %{{.*}})
+  vul = vec_gfmsum_accum(vui, vui, vul);
+  // CHECK: call <2 x i64> @llvm.s390.vgfmaf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <2 x i64> %{{.*}})
+  vuc = vec_gfmsum_accum_128(vul, vul, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vgfmag(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <16 x i8> %{{.*}})
+
+  vsc = vec_mladd(vsc, vsc, vsc);
+  vsc = vec_mladd(vuc, vsc, vsc);
+  vsc = vec_mladd(vsc, vuc, vuc);
+  vuc = vec_mladd(vuc, vuc, vuc);
+  vss = vec_mladd(vss, vss, vss);
+  vss = vec_mladd(vus, vss, vss);
+  vss = vec_mladd(vss, vus, vus);
+  vus = vec_mladd(vus, vus, vus);
+  vsi = vec_mladd(vsi, vsi, vsi);
+  vsi = vec_mladd(vui, vsi, vsi);
+  vsi = vec_mladd(vsi, vui, vui);
+  vui = vec_mladd(vui, vui, vui);
+
+  vsc = vec_mhadd(vsc, vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vmahb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_mhadd(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vmalhb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_mhadd(vss, vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vmahh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_mhadd(vus, vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vmalhh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_mhadd(vsi, vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vmahf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_mhadd(vui, vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vmalhf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vss = vec_meadd(vsc, vsc, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vmaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_meadd(vuc, vuc, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vmaleb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_meadd(vss, vss, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vmaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_meadd(vus, vus, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vmaleh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <4 x i32> %{{.*}})
+  vsl = vec_meadd(vsi, vsi, vsl);
+  // CHECK: call <2 x i64> @llvm.s390.vmaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <2 x i64> %{{.*}})
+  vul = vec_meadd(vui, vui, vul);
+  // CHECK: call <2 x i64> @llvm.s390.vmalef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <2 x i64> %{{.*}})
+
+  vss = vec_moadd(vsc, vsc, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vmaob(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_moadd(vuc, vuc, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vmalob(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_moadd(vss, vss, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vmaoh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_moadd(vus, vus, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vmaloh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <4 x i32> %{{.*}})
+  vsl = vec_moadd(vsi, vsi, vsl);
+  // CHECK: call <2 x i64> @llvm.s390.vmaof(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <2 x i64> %{{.*}})
+  vul = vec_moadd(vui, vui, vul);
+  // CHECK: call <2 x i64> @llvm.s390.vmalof(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <2 x i64> %{{.*}})
+
+  vsc = vec_mulh(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vmhb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_mulh(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vmlhb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_mulh(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vmhh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_mulh(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vmlhh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_mulh(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vmhf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_mulh(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vmlhf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vss = vec_mule(vsc, vsc);
+  // CHECK: call <8 x i16> @llvm.s390.vmeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_mule(vuc, vuc);
+  // CHECK: call <8 x i16> @llvm.s390.vmleb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_mule(vss, vss);
+  // CHECK: call <4 x i32> @llvm.s390.vmeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vui = vec_mule(vus, vus);
+  // CHECK: call <4 x i32> @llvm.s390.vmleh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsl = vec_mule(vsi, vsi);
+  // CHECK: call <2 x i64> @llvm.s390.vmef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vul = vec_mule(vui, vui);
+  // CHECK: call <2 x i64> @llvm.s390.vmlef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vss = vec_mulo(vsc, vsc);
+  // CHECK: call <8 x i16> @llvm.s390.vmob(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_mulo(vuc, vuc);
+  // CHECK: call <8 x i16> @llvm.s390.vmlob(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vsi = vec_mulo(vss, vss);
+  // CHECK: call <4 x i32> @llvm.s390.vmoh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vui = vec_mulo(vus, vus);
+  // CHECK: call <4 x i32> @llvm.s390.vmloh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsl = vec_mulo(vsi, vsi);
+  // CHECK: call <2 x i64> @llvm.s390.vmof(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vul = vec_mulo(vui, vui);
+  // CHECK: call <2 x i64> @llvm.s390.vmlof(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vuc = vec_subc(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vscbib(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vus = vec_subc(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vscbih(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vui = vec_subc(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vscbif(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vul = vec_subc(vul, vul);
+  // CHECK: call <2 x i64> @llvm.s390.vscbig(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  vuc = vec_sub_u128(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsq(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_subc_u128(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vscbiq(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_sube_u128(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsbiq(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_subec_u128(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vsbcbiq(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+
+  vui = vec_sum4(vuc, vuc);
+  // CHECK: call <4 x i32> @llvm.s390.vsumb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vui = vec_sum4(vus, vus);
+  // CHECK: call <4 x i32> @llvm.s390.vsumh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vul = vec_sum2(vus, vus);
+  // CHECK: call <2 x i64> @llvm.s390.vsumgh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vul = vec_sum2(vui, vui);
+  // CHECK: call <2 x i64> @llvm.s390.vsumgf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vuc = vec_sum_u128(vui, vui);
+  // CHECK: call <16 x i8> @llvm.s390.vsumqf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vuc = vec_sum_u128(vul, vul);
+  // CHECK: call <16 x i8> @llvm.s390.vsumqg(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
+
+  idx = vec_test_mask(vsc, vuc);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_test_mask(vuc, vuc);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_test_mask(vss, vus);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_test_mask(vus, vus);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_test_mask(vsi, vui);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_test_mask(vui, vui);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_test_mask(vsl, vul);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_test_mask(vul, vul);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  idx = vec_test_mask(vd, vul);
+  // CHECK: call i32 @llvm.s390.vtm(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+}
+
+void test_string(void) {
+  vsc = vec_cp_until_zero(vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vistrb(<16 x i8> %{{.*}})
+  vuc = vec_cp_until_zero(vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vistrb(<16 x i8> %{{.*}})
+  vbc = vec_cp_until_zero(vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vistrb(<16 x i8> %{{.*}})
+  vss = vec_cp_until_zero(vss);
+  // CHECK: call <8 x i16> @llvm.s390.vistrh(<8 x i16> %{{.*}})
+  vus = vec_cp_until_zero(vus);
+  // CHECK: call <8 x i16> @llvm.s390.vistrh(<8 x i16> %{{.*}})
+  vbs = vec_cp_until_zero(vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vistrh(<8 x i16> %{{.*}})
+  vsi = vec_cp_until_zero(vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vistrf(<4 x i32> %{{.*}})
+  vui = vec_cp_until_zero(vui);
+  // CHECK: call <4 x i32> @llvm.s390.vistrf(<4 x i32> %{{.*}})
+  vbi = vec_cp_until_zero(vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vistrf(<4 x i32> %{{.*}})
+
+  vsc = vec_cp_until_zero_cc(vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vistrbs(<16 x i8> %{{.*}})
+  vuc = vec_cp_until_zero_cc(vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vistrbs(<16 x i8> %{{.*}})
+  vbc = vec_cp_until_zero_cc(vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vistrbs(<16 x i8> %{{.*}})
+  vss = vec_cp_until_zero_cc(vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vistrhs(<8 x i16> %{{.*}})
+  vus = vec_cp_until_zero_cc(vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vistrhs(<8 x i16> %{{.*}})
+  vbs = vec_cp_until_zero_cc(vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vistrhs(<8 x i16> %{{.*}})
+  vsi = vec_cp_until_zero_cc(vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vistrfs(<4 x i32> %{{.*}})
+  vui = vec_cp_until_zero_cc(vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vistrfs(<4 x i32> %{{.*}})
+  vbi = vec_cp_until_zero_cc(vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vistrfs(<4 x i32> %{{.*}})
+
+  vsc = vec_cmpeq_idx(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpeq_idx(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpeq_idx(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_cmpeq_idx(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfeeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpeq_idx(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfeeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpeq_idx(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfeeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_cmpeq_idx(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfeef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpeq_idx(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfeef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpeq_idx(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfeef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vsc = vec_cmpeq_idx_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfeebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpeq_idx_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfeebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpeq_idx_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfeebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_cmpeq_idx_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfeehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpeq_idx_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfeehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpeq_idx_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfeehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_cmpeq_idx_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfeefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpeq_idx_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfeefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpeq_idx_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfeefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vsc = vec_cmpeq_or_0_idx(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpeq_or_0_idx(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpeq_or_0_idx(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_cmpeq_or_0_idx(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfeezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpeq_or_0_idx(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfeezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpeq_or_0_idx(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfeezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_cmpeq_or_0_idx(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfeezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpeq_or_0_idx(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfeezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpeq_or_0_idx(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfeezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vsc = vec_cmpeq_or_0_idx_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfeezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpeq_or_0_idx_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfeezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpeq_or_0_idx_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfeezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_cmpeq_or_0_idx_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfeezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpeq_or_0_idx_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfeezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpeq_or_0_idx_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfeezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_cmpeq_or_0_idx_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfeezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpeq_or_0_idx_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfeezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpeq_or_0_idx_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfeezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vsc = vec_cmpne_idx(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeneb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpne_idx(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeneb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpne_idx(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfeneb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_cmpne_idx(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfeneh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpne_idx(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfeneh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpne_idx(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfeneh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_cmpne_idx(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfenef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpne_idx(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfenef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpne_idx(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfenef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vsc = vec_cmpne_idx_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfenebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpne_idx_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfenebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpne_idx_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfenebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_cmpne_idx_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfenehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpne_idx_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfenehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpne_idx_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfenehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_cmpne_idx_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfenefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpne_idx_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfenefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpne_idx_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfenefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vsc = vec_cmpne_or_0_idx(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfenezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpne_or_0_idx(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfenezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpne_or_0_idx(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfenezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_cmpne_or_0_idx(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfenezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpne_or_0_idx(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfenezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpne_or_0_idx(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfenezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_cmpne_or_0_idx(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfenezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpne_or_0_idx(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfenezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpne_or_0_idx(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfenezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vsc = vec_cmpne_or_0_idx_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfenezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpne_or_0_idx_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfenezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vuc = vec_cmpne_or_0_idx_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfenezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
+  vss = vec_cmpne_or_0_idx_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfenezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpne_or_0_idx_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfenezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vus = vec_cmpne_or_0_idx_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfenezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
+  vsi = vec_cmpne_or_0_idx_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfenezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpne_or_0_idx_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfenezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+  vui = vec_cmpne_or_0_idx_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfenezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
+
+  vbc = vec_cmprg(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vstrcb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 4)
+  vbs = vec_cmprg(vus, vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vstrch(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 4)
+  vbi = vec_cmprg(vui, vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vstrcf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 4)
+
+  vbc = vec_cmprg_cc(vuc, vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vstrcbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 4)
+  vbs = vec_cmprg_cc(vus, vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vstrchs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 4)
+  vbi = vec_cmprg_cc(vui, vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vstrcfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 4)
+
+  vuc = vec_cmprg_idx(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vstrcb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vus = vec_cmprg_idx(vus, vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vstrch(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vui = vec_cmprg_idx(vui, vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vstrcf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+
+  vuc = vec_cmprg_idx_cc(vuc, vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vstrcbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vus = vec_cmprg_idx_cc(vus, vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vstrchs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vui = vec_cmprg_idx_cc(vui, vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vstrcfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+
+  vuc = vec_cmprg_or_0_idx(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vstrczb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vus = vec_cmprg_or_0_idx(vus, vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vstrczh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vui = vec_cmprg_or_0_idx(vui, vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vstrczf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+
+  vuc = vec_cmprg_or_0_idx_cc(vuc, vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vstrczbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vus = vec_cmprg_or_0_idx_cc(vus, vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vstrczhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vui = vec_cmprg_or_0_idx_cc(vui, vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vstrczfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+
+  vbc = vec_cmpnrg(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vstrcb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vbs = vec_cmpnrg(vus, vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vstrch(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 12)
+  vbi = vec_cmpnrg(vui, vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vstrcf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 12)
+
+  vbc = vec_cmpnrg_cc(vuc, vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vstrcbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vbs = vec_cmpnrg_cc(vus, vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vstrchs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 12)
+  vbi = vec_cmpnrg_cc(vui, vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vstrcfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 12)
+
+  vuc = vec_cmpnrg_idx(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vstrcb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vus = vec_cmpnrg_idx(vus, vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vstrch(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vui = vec_cmpnrg_idx(vui, vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vstrcf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+
+  vuc = vec_cmpnrg_idx_cc(vuc, vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vstrcbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vus = vec_cmpnrg_idx_cc(vus, vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vstrchs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vui = vec_cmpnrg_idx_cc(vui, vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vstrcfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+
+  vuc = vec_cmpnrg_or_0_idx(vuc, vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vstrczb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vus = vec_cmpnrg_or_0_idx(vus, vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vstrczh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vui = vec_cmpnrg_or_0_idx(vui, vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vstrczf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+
+  vuc = vec_cmpnrg_or_0_idx_cc(vuc, vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vstrczbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vus = vec_cmpnrg_or_0_idx_cc(vus, vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vstrczhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vui = vec_cmpnrg_or_0_idx_cc(vui, vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vstrczfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+
+  vbc = vec_find_any_eq(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 4)
+  vbc = vec_find_any_eq(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 4)
+  vbc = vec_find_any_eq(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 4)
+  vbs = vec_find_any_eq(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 4)
+  vbs = vec_find_any_eq(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 4)
+  vbs = vec_find_any_eq(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 4)
+  vbi = vec_find_any_eq(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 4)
+  vbi = vec_find_any_eq(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 4)
+  vbi = vec_find_any_eq(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 4)
+
+  vbc = vec_find_any_eq_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 4)
+  vbc = vec_find_any_eq_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 4)
+  vbc = vec_find_any_eq_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 4)
+  vbs = vec_find_any_eq_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 4)
+  vbs = vec_find_any_eq_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 4)
+  vbs = vec_find_any_eq_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 4)
+  vbi = vec_find_any_eq_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 4)
+  vbi = vec_find_any_eq_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 4)
+  vbi = vec_find_any_eq_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 4)
+
+  vsc = vec_find_any_eq_idx(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_find_any_eq_idx(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_find_any_eq_idx(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vss = vec_find_any_eq_idx(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_find_any_eq_idx(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_find_any_eq_idx(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vsi = vec_find_any_eq_idx(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_find_any_eq_idx(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_find_any_eq_idx(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+
+  vsc = vec_find_any_eq_idx_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_find_any_eq_idx_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_find_any_eq_idx_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vss = vec_find_any_eq_idx_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_find_any_eq_idx_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_find_any_eq_idx_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vsi = vec_find_any_eq_idx_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_find_any_eq_idx_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_find_any_eq_idx_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+
+  vsc = vec_find_any_eq_or_0_idx(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_find_any_eq_or_0_idx(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_find_any_eq_or_0_idx(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vss = vec_find_any_eq_or_0_idx(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfaezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_find_any_eq_or_0_idx(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfaezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_find_any_eq_or_0_idx(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfaezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vsi = vec_find_any_eq_or_0_idx(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_find_any_eq_or_0_idx(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfaezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_find_any_eq_or_0_idx(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+
+  vsc = vec_find_any_eq_or_0_idx_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_find_any_eq_or_0_idx_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vuc = vec_find_any_eq_or_0_idx_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 0)
+  vss = vec_find_any_eq_or_0_idx_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_find_any_eq_or_0_idx_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vus = vec_find_any_eq_or_0_idx_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 0)
+  vsi = vec_find_any_eq_or_0_idx_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_find_any_eq_or_0_idx_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+  vui = vec_find_any_eq_or_0_idx_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 0)
+
+  vbc = vec_find_any_ne(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vbc = vec_find_any_ne(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vbc = vec_find_any_ne(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vbs = vec_find_any_ne(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 12)
+  vbs = vec_find_any_ne(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 12)
+  vbs = vec_find_any_ne(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 12)
+  vbi = vec_find_any_ne(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 12)
+  vbi = vec_find_any_ne(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 12)
+  vbi = vec_find_any_ne(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 12)
+
+  vbc = vec_find_any_ne_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vbc = vec_find_any_ne_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vbc = vec_find_any_ne_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 12)
+  vbs = vec_find_any_ne_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 12)
+  vbs = vec_find_any_ne_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 12)
+  vbs = vec_find_any_ne_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 12)
+  vbi = vec_find_any_ne_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 12)
+  vbi = vec_find_any_ne_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 12)
+  vbi = vec_find_any_ne_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 12)
+
+  vsc = vec_find_any_ne_idx(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vuc = vec_find_any_ne_idx(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vuc = vec_find_any_ne_idx(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaeb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vss = vec_find_any_ne_idx(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vus = vec_find_any_ne_idx(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vus = vec_find_any_ne_idx(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfaeh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vsi = vec_find_any_ne_idx(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+  vui = vec_find_any_ne_idx(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+  vui = vec_find_any_ne_idx(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaef(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+
+  vsc = vec_find_any_ne_idx_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vuc = vec_find_any_ne_idx_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vuc = vec_find_any_ne_idx_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaebs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vss = vec_find_any_ne_idx_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vus = vec_find_any_ne_idx_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vus = vec_find_any_ne_idx_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaehs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vsi = vec_find_any_ne_idx_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+  vui = vec_find_any_ne_idx_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+  vui = vec_find_any_ne_idx_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaefs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+
+  vsc = vec_find_any_ne_or_0_idx(vsc, vsc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vuc = vec_find_any_ne_or_0_idx(vuc, vuc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vuc = vec_find_any_ne_or_0_idx(vbc, vbc);
+  // CHECK: call <16 x i8> @llvm.s390.vfaezb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vss = vec_find_any_ne_or_0_idx(vss, vss);
+  // CHECK: call <8 x i16> @llvm.s390.vfaezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vus = vec_find_any_ne_or_0_idx(vus, vus);
+  // CHECK: call <8 x i16> @llvm.s390.vfaezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vus = vec_find_any_ne_or_0_idx(vbs, vbs);
+  // CHECK: call <8 x i16> @llvm.s390.vfaezh(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vsi = vec_find_any_ne_or_0_idx(vsi, vsi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+  vui = vec_find_any_ne_or_0_idx(vui, vui);
+  // CHECK: call <4 x i32> @llvm.s390.vfaezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+  vui = vec_find_any_ne_or_0_idx(vbi, vbi);
+  // CHECK: call <4 x i32> @llvm.s390.vfaezf(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+
+  vsc = vec_find_any_ne_or_0_idx_cc(vsc, vsc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vuc = vec_find_any_ne_or_0_idx_cc(vuc, vuc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vuc = vec_find_any_ne_or_0_idx_cc(vbc, vbc, &cc);
+  // CHECK: call { <16 x i8>, i32 } @llvm.s390.vfaezbs(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i32 8)
+  vss = vec_find_any_ne_or_0_idx_cc(vss, vss, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vus = vec_find_any_ne_or_0_idx_cc(vus, vus, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vus = vec_find_any_ne_or_0_idx_cc(vbs, vbs, &cc);
+  // CHECK: call { <8 x i16>, i32 } @llvm.s390.vfaezhs(<8 x i16> %{{.*}}, <8 x i16> %{{.*}}, i32 8)
+  vsi = vec_find_any_ne_or_0_idx_cc(vsi, vsi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+  vui = vec_find_any_ne_or_0_idx_cc(vui, vui, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+  vui = vec_find_any_ne_or_0_idx_cc(vbi, vbi, &cc);
+  // CHECK: call { <4 x i32>, i32 } @llvm.s390.vfaezfs(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}, i32 8)
+}
+
+void test_float(void) {
+  vd = vec_abs(vd);
+  // CHECK: call <2 x double> @llvm.fabs.v2f64(<2 x double> %{{.*}})
+
+  vd = vec_nabs(vd);
+  // CHECK: [[ABS:%[^ ]+]] = tail call <2 x double> @llvm.fabs.v2f64(<2 x double> %{{.*}})
+  // CHECK-NEXT: fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, [[ABS]]
+
+  vd = vec_madd(vd, vd, vd);
+  // CHECK: call <2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}})
+  vd = vec_msub(vd, vd, vd);
+  // CHECK: [[NEG:%[^ ]+]] = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %{{.*}}
+  // CHECK: call <2 x double> @llvm.fma.v2f64(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> [[NEG]])
+  vd = vec_sqrt(vd);
+  // CHECK: call <2 x double> @llvm.sqrt.v2f64(<2 x double> %{{.*}})
+
+  vd = vec_ld2f(cptrf);
+  // CHECK: [[VAL:%[^ ]+]] = load <2 x float>, <2 x float>* %{{.*}}
+  // CHECK: fpext <2 x float> [[VAL]] to <2 x double>
+  vec_st2f(vd, ptrf);
+  // CHECK: [[VAL:%[^ ]+]] = fptrunc <2 x double> %{{.*}} to <2 x float>
+  // CHECK: store <2 x float> [[VAL]], <2 x float>* %{{.*}}
+
+  vd = vec_ctd(vsl, 0);
+  // CHECK: sitofp <2 x i64> %{{.*}} to <2 x double>
+  vd = vec_ctd(vul, 0);
+  // CHECK: uitofp <2 x i64> %{{.*}} to <2 x double>
+  vd = vec_ctd(vsl, 1);
+  // CHECK: [[VAL:%[^ ]+]] = sitofp <2 x i64> %{{.*}} to <2 x double>
+  // CHECK: fmul <2 x double> [[VAL]], <double 5.000000e-01, double 5.000000e-01>
+  vd = vec_ctd(vul, 1);
+  // CHECK: [[VAL:%[^ ]+]] = uitofp <2 x i64> %{{.*}} to <2 x double>
+  // CHECK: fmul <2 x double> [[VAL]], <double 5.000000e-01, double 5.000000e-01>
+  vd = vec_ctd(vsl, 31);
+  // CHECK: [[VAL:%[^ ]+]] = sitofp <2 x i64> %{{.*}} to <2 x double>
+  // CHECK: fmul <2 x double> [[VAL]], <double 0x3E00000000000000, double 0x3E00000000000000>
+  vd = vec_ctd(vul, 31);
+  // CHECK: [[VAL:%[^ ]+]] = uitofp <2 x i64> %{{.*}} to <2 x double>
+  // CHECK: fmul <2 x double> [[VAL]], <double 0x3E00000000000000, double 0x3E00000000000000>
+
+  vsl = vec_ctsl(vd, 0);
+  // CHECK: fptosi <2 x double> %{{.*}} to <2 x i64>
+  vul = vec_ctul(vd, 0);
+  // CHECK: fptoui <2 x double> %{{.*}} to <2 x i64>
+  vsl = vec_ctsl(vd, 1);
+  // CHECK: [[VAL:%[^ ]+]] = fmul <2 x double> %{{.*}}, <double 2.000000e+00, double 2.000000e+00>
+  // CHECK: fptosi <2 x double> [[VAL]] to <2 x i64>
+  vul = vec_ctul(vd, 1);
+  // CHECK: [[VAL:%[^ ]+]] = fmul <2 x double> %{{.*}}, <double 2.000000e+00, double 2.000000e+00>
+  // CHECK: fptoui <2 x double> [[VAL]] to <2 x i64>
+  vsl = vec_ctsl(vd, 31);
+  // CHECK: [[VAL:%[^ ]+]] = fmul <2 x double> %{{.*}}, <double 0x41E0000000000000, double 0x41E0000000000000>
+  // CHECK: fptosi <2 x double> [[VAL]] to <2 x i64>
+  vul = vec_ctul(vd, 31);
+  // CHECK: [[VAL:%[^ ]+]] = fmul <2 x double> %{{.*}}, <double 0x41E0000000000000, double 0x41E0000000000000>
+  // CHECK: fptoui <2 x double> [[VAL]] to <2 x i64>
+
+  vd = vec_roundp(vd);
+  // CHECK: call <2 x double> @llvm.ceil.v2f64(<2 x double> %{{.*}})
+  vd = vec_ceil(vd);
+  // CHECK: call <2 x double> @llvm.ceil.v2f64(<2 x double> %{{.*}})
+  vd = vec_roundm(vd);
+  // CHECK: call <2 x double> @llvm.floor.v2f64(<2 x double> %{{.*}})
+  vd = vec_floor(vd);
+  // CHECK: call <2 x double> @llvm.floor.v2f64(<2 x double> %{{.*}})
+  vd = vec_roundz(vd);
+  // CHECK: call <2 x double> @llvm.trunc.v2f64(<2 x double> %{{.*}})
+  vd = vec_trunc(vd);
+  // CHECK: call <2 x double> @llvm.trunc.v2f64(<2 x double> %{{.*}})
+  vd = vec_roundc(vd);
+  // CHECK: call <2 x double> @llvm.nearbyint.v2f64(<2 x double> %{{.*}})
+  vd = vec_round(vd);
+  // CHECK: call <2 x double> @llvm.s390.vfidb(<2 x double> %{{.*}}, i32 4, i32 4)
+
+  vbl = vec_fp_test_data_class(vd, 0, &cc);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vftcidb(<2 x double> %{{.*}}, i32 0)
+  vbl = vec_fp_test_data_class(vd, 4095, &cc);
+  // CHECK: call { <2 x i64>, i32 } @llvm.s390.vftcidb(<2 x double> %{{.*}}, i32 4095)
+}
diff --git a/clang/test/CodeGen/zvector.c b/clang/test/CodeGen/zvector.c
new file mode 100644
index 0000000..ebe7e41
--- /dev/null
+++ b/clang/test/CodeGen/zvector.c
@@ -0,0 +1,2798 @@
+// RUN: %clang_cc1 -triple s390x-linux-gnu -target-cpu z13 -fzvector \
+// RUN:  -O -emit-llvm -o - -W -Wall -Werror %s | FileCheck %s
+
+volatile vector signed char sc, sc2;
+volatile vector unsigned char uc, uc2;
+volatile vector bool char bc, bc2;
+
+volatile vector signed short ss, ss2;
+volatile vector unsigned short us, us2;
+volatile vector bool short bs, bs2;
+
+volatile vector signed int si, si2;
+volatile vector unsigned int ui, ui2;
+volatile vector bool int bi, bi2;
+
+volatile vector signed long long sl, sl2;
+volatile vector unsigned long long ul, ul2;
+volatile vector bool long long bl, bl2;
+
+volatile vector double fd, fd2;
+
+volatile int cnt;
+
+void test_assign (void)
+{
+// CHECK-LABEL: test_assign
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: store volatile <16 x i8> [[VAL]], <16 x i8>* @sc
+  sc = sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: store volatile <16 x i8> [[VAL]], <16 x i8>* @uc
+  uc = uc2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: store volatile <8 x i16> [[VAL]], <8 x i16>* @ss
+  ss = ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: store volatile <8 x i16> [[VAL]], <8 x i16>* @us
+  us = us2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: store volatile <4 x i32> [[VAL]], <4 x i32>* @si
+  si = si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: store volatile <4 x i32> [[VAL]], <4 x i32>* @ui
+  ui = ui2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: store volatile <2 x i64> [[VAL]], <2 x i64>* @sl
+  sl = sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: store volatile <2 x i64> [[VAL]], <2 x i64>* @ul
+  ul = ul2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: store volatile <2 x double> [[VAL]], <2 x double>* @fd
+  fd = fd2;
+}
+
+void test_pos (void)
+{
+// CHECK-LABEL: test_pos
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: store volatile <16 x i8> [[VAL]], <16 x i8>* @sc
+  sc = +sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: store volatile <16 x i8> [[VAL]], <16 x i8>* @uc
+  uc = +uc2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: store volatile <8 x i16> [[VAL]], <8 x i16>* @ss
+  ss = +ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: store volatile <8 x i16> [[VAL]], <8 x i16>* @us
+  us = +us2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: store volatile <4 x i32> [[VAL]], <4 x i32>* @si
+  si = +si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: store volatile <4 x i32> [[VAL]], <4 x i32>* @ui
+  ui = +ui2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: store volatile <2 x i64> [[VAL]], <2 x i64>* @sl
+  sl = +sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: store volatile <2 x i64> [[VAL]], <2 x i64>* @ul
+  ul = +ul2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: store volatile <2 x double> [[VAL]], <2 x double>* @fd
+  fd = +fd2;
+}
+
+void test_neg (void)
+{
+// CHECK-LABEL: test_neg
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = sub <16 x i8> zeroinitializer, [[VAL]]
+  sc = -sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = sub <8 x i16> zeroinitializer, [[VAL]]
+  ss = -ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = sub <4 x i32> zeroinitializer, [[VAL]]
+  si = -si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = sub <2 x i64> zeroinitializer, [[VAL]]
+  sl = -sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, [[VAL]]
+  fd = -fd2;
+}
+
+void test_preinc (void)
+{
+// CHECK-LABEL: test_preinc
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL]], <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
+  ++sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL]], <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
+  ++uc2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL]], <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
+  ++ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL]], <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
+  ++us2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL]], <i32 1, i32 1, i32 1, i32 1>
+  ++si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL]], <i32 1, i32 1, i32 1, i32 1>
+  ++ui2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL]], <i64 1, i64 1>
+  ++sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL]], <i64 1, i64 1>
+  ++ul2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fadd <2 x double> [[VAL]], <double 1.000000e+00, double 1.000000e+00>
+  ++fd2;
+}
+
+void test_postinc (void)
+{
+// CHECK-LABEL: test_postinc
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL]], <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
+  sc2++;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL]], <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
+  uc2++;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL]], <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
+  ss2++;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL]], <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
+  us2++;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL]], <i32 1, i32 1, i32 1, i32 1>
+  si2++;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL]], <i32 1, i32 1, i32 1, i32 1>
+  ui2++;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL]], <i64 1, i64 1>
+  sl2++;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL]], <i64 1, i64 1>
+  ul2++;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fadd <2 x double> [[VAL]], <double 1.000000e+00, double 1.000000e+00>
+  fd2++;
+}
+
+void test_predec (void)
+{
+// CHECK-LABEL: test_predec
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL]], <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+  --sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL]], <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+  --uc2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL]], <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+  --ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL]], <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+  --us2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL]], <i32 -1, i32 -1, i32 -1, i32 -1>
+  --si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL]], <i32 -1, i32 -1, i32 -1, i32 -1>
+  --ui2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL]], <i64 -1, i64 -1>
+  --sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL]], <i64 -1, i64 -1>
+  --ul2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fadd <2 x double> [[VAL]], <double -1.000000e+00, double -1.000000e+00>
+  --fd2;
+}
+
+void test_postdec (void)
+{
+// CHECK-LABEL: test_postdec
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL]], <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+  sc2--;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL]], <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+  uc2--;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL]], <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+  ss2--;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL]], <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+  us2--;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL]], <i32 -1, i32 -1, i32 -1, i32 -1>
+  si2--;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL]], <i32 -1, i32 -1, i32 -1, i32 -1>
+  ui2--;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL]], <i64 -1, i64 -1>
+  sl2--;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL]], <i64 -1, i64 -1>
+  ul2--;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fadd <2 x double> [[VAL]], <double -1.000000e+00, double -1.000000e+00>
+  fd2--;
+}
+
+void test_add (void)
+{
+// CHECK-LABEL: test_add
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL2]], [[VAL1]]
+  sc = sc + sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL2]], [[VAL1]]
+  sc = sc + bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL2]], [[VAL1]]
+  sc = bc + sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL2]], [[VAL1]]
+  uc = uc + uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL2]], [[VAL1]]
+  uc = uc + bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = add <16 x i8> [[VAL2]], [[VAL1]]
+  uc = bc + uc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL2]], [[VAL1]]
+  ss = ss + ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL2]], [[VAL1]]
+  ss = ss + bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL2]], [[VAL1]]
+  ss = bs + ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL2]], [[VAL1]]
+  us = us + us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL2]], [[VAL1]]
+  us = us + bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = add <8 x i16> [[VAL2]], [[VAL1]]
+  us = bs + us2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL2]], [[VAL1]]
+  si = si + si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL2]], [[VAL1]]
+  si = si + bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL2]], [[VAL1]]
+  si = bi + si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL2]], [[VAL1]]
+  ui = ui + ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL2]], [[VAL1]]
+  ui = ui + bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = add <4 x i32> [[VAL2]], [[VAL1]]
+  ui = bi + ui2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL2]], [[VAL1]]
+  sl = sl + sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL2]], [[VAL1]]
+  sl = sl + bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL2]], [[VAL1]]
+  sl = bl + sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL2]], [[VAL1]]
+  ul = ul + ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL2]], [[VAL1]]
+  ul = ul + bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = add <2 x i64> [[VAL2]], [[VAL1]]
+  ul = bl + ul2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fadd <2 x double> [[VAL1]], [[VAL2]]
+  fd = fd + fd2;
+}
+
+void test_add_assign (void)
+{
+// CHECK-LABEL: test_add_assign
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = add <16 x i8> [[VAL1]], [[VAL2]]
+  sc += sc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = add <16 x i8> [[VAL1]], [[VAL2]]
+  sc += bc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = add <16 x i8> [[VAL1]], [[VAL2]]
+  uc += uc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = add <16 x i8> [[VAL1]], [[VAL2]]
+  uc += bc2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = add <8 x i16> [[VAL1]], [[VAL2]]
+  ss += ss2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = add <8 x i16> [[VAL1]], [[VAL2]]
+  ss += bs2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = add <8 x i16> [[VAL1]], [[VAL2]]
+  us += us2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = add <8 x i16> [[VAL1]], [[VAL2]]
+  us += bs2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = add <4 x i32> [[VAL1]], [[VAL2]]
+  si += si2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = add <4 x i32> [[VAL1]], [[VAL2]]
+  si += bi2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = add <4 x i32> [[VAL1]], [[VAL2]]
+  ui += ui2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = add <4 x i32> [[VAL1]], [[VAL2]]
+  ui += bi2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = add <2 x i64> [[VAL1]], [[VAL2]]
+  sl += sl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = add <2 x i64> [[VAL1]], [[VAL2]]
+  sl += bl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = add <2 x i64> [[VAL1]], [[VAL2]]
+  ul += ul2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = add <2 x i64> [[VAL1]], [[VAL2]]
+  ul += bl2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: %{{.*}} = fadd <2 x double> [[VAL2]], [[VAL1]]
+  fd += fd2;
+}
+
+void test_sub (void)
+{
+// CHECK-LABEL: test_sub
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  sc = sc - sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  sc = sc - bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  sc = bc - sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  uc = uc - uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  uc = uc - bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  uc = bc - uc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  ss = ss - ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  ss = ss - bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  ss = bs - ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  us = us - us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  us = us - bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  us = bs - us2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  si = si - si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  si = si - bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  si = bi - si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  ui = ui - ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  ui = ui - bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  ui = bi - ui2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  sl = sl - sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  sl = sl - bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  sl = bl - sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  ul = ul - ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  ul = ul - bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  ul = bl - ul2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fsub <2 x double> [[VAL1]], [[VAL2]]
+  fd = fd - fd2;
+}
+
+void test_sub_assign (void)
+{
+// CHECK-LABEL: test_sub_assign
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  sc -= sc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  sc -= bc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  uc -= uc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = sub <16 x i8> [[VAL1]], [[VAL2]]
+  uc -= bc2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  ss -= ss2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  ss -= bs2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  us -= us2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = sub <8 x i16> [[VAL1]], [[VAL2]]
+  us -= bs2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  si -= si2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  si -= bi2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  ui -= ui2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = sub <4 x i32> [[VAL1]], [[VAL2]]
+  ui -= bi2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  sl -= sl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  sl -= bl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  ul -= ul2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = sub <2 x i64> [[VAL1]], [[VAL2]]
+  ul -= bl2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: %{{.*}} = fsub <2 x double> [[VAL1]], [[VAL2]]
+  fd -= fd2;
+}
+
+void test_mul (void)
+{
+// CHECK-LABEL: test_mul
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = mul <16 x i8> [[VAL2]], [[VAL1]]
+  sc = sc * sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = mul <16 x i8> [[VAL2]], [[VAL1]]
+  uc = uc * uc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = mul <8 x i16> [[VAL2]], [[VAL1]]
+  ss = ss * ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = mul <8 x i16> [[VAL2]], [[VAL1]]
+  us = us * us2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = mul <4 x i32> [[VAL2]], [[VAL1]]
+  si = si * si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = mul <4 x i32> [[VAL2]], [[VAL1]]
+  ui = ui * ui2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = mul <2 x i64> [[VAL2]], [[VAL1]]
+  sl = sl * sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = mul <2 x i64> [[VAL2]], [[VAL1]]
+  ul = ul * ul2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fmul <2 x double> [[VAL1]], [[VAL2]]
+  fd = fd * fd2;
+}
+
+void test_mul_assign (void)
+{
+// CHECK-LABEL: test_mul_assign
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = mul <16 x i8> [[VAL1]], [[VAL2]]
+  sc *= sc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = mul <16 x i8> [[VAL1]], [[VAL2]]
+  uc *= uc2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = mul <8 x i16> [[VAL1]], [[VAL2]]
+  ss *= ss2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = mul <8 x i16> [[VAL1]], [[VAL2]]
+  us *= us2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = mul <4 x i32> [[VAL1]], [[VAL2]]
+  si *= si2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = mul <4 x i32> [[VAL1]], [[VAL2]]
+  ui *= ui2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = mul <2 x i64> [[VAL1]], [[VAL2]]
+  sl *= sl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = mul <2 x i64> [[VAL1]], [[VAL2]]
+  ul *= ul2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: %{{.*}} = fmul <2 x double> [[VAL2]], [[VAL1]]
+  fd *= fd2;
+}
+
+void test_div (void)
+{
+// CHECK-LABEL: test_div
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = sdiv <16 x i8> [[VAL1]], [[VAL2]]
+  sc = sc / sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = udiv <16 x i8> [[VAL1]], [[VAL2]]
+  uc = uc / uc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = sdiv <8 x i16> [[VAL1]], [[VAL2]]
+  ss = ss / ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = udiv <8 x i16> [[VAL1]], [[VAL2]]
+  us = us / us2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = sdiv <4 x i32> [[VAL1]], [[VAL2]]
+  si = si / si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = udiv <4 x i32> [[VAL1]], [[VAL2]]
+  ui = ui / ui2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = sdiv <2 x i64> [[VAL1]], [[VAL2]]
+  sl = sl / sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = udiv <2 x i64> [[VAL1]], [[VAL2]]
+  ul = ul / ul2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: %{{.*}} = fdiv <2 x double> [[VAL1]], [[VAL2]]
+  fd = fd / fd2;
+}
+
+void test_div_assign (void)
+{
+// CHECK-LABEL: test_div_assign
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = sdiv <16 x i8> [[VAL1]], [[VAL2]]
+  sc /= sc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = udiv <16 x i8> [[VAL1]], [[VAL2]]
+  uc /= uc2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = sdiv <8 x i16> [[VAL1]], [[VAL2]]
+  ss /= ss2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = udiv <8 x i16> [[VAL1]], [[VAL2]]
+  us /= us2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = sdiv <4 x i32> [[VAL1]], [[VAL2]]
+  si /= si2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = udiv <4 x i32> [[VAL1]], [[VAL2]]
+  ui /= ui2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = sdiv <2 x i64> [[VAL1]], [[VAL2]]
+  sl /= sl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = udiv <2 x i64> [[VAL1]], [[VAL2]]
+  ul /= ul2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: %{{.*}} = fdiv <2 x double> [[VAL1]], [[VAL2]]
+  fd /= fd2;
+}
+
+void test_rem (void)
+{
+// CHECK-LABEL: test_rem
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = srem <16 x i8> [[VAL1]], [[VAL2]]
+  sc = sc % sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = urem <16 x i8> [[VAL1]], [[VAL2]]
+  uc = uc % uc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = srem <8 x i16> [[VAL1]], [[VAL2]]
+  ss = ss % ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = urem <8 x i16> [[VAL1]], [[VAL2]]
+  us = us % us2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = srem <4 x i32> [[VAL1]], [[VAL2]]
+  si = si % si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = urem <4 x i32> [[VAL1]], [[VAL2]]
+  ui = ui % ui2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = srem <2 x i64> [[VAL1]], [[VAL2]]
+  sl = sl % sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = urem <2 x i64> [[VAL1]], [[VAL2]]
+  ul = ul % ul2;
+}
+
+void test_rem_assign (void)
+{
+// CHECK-LABEL: test_rem_assign
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = srem <16 x i8> [[VAL1]], [[VAL2]]
+  sc %= sc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = urem <16 x i8> [[VAL1]], [[VAL2]]
+  uc %= uc2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = srem <8 x i16> [[VAL1]], [[VAL2]]
+  ss %= ss2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = urem <8 x i16> [[VAL1]], [[VAL2]]
+  us %= us2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = srem <4 x i32> [[VAL1]], [[VAL2]]
+  si %= si2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = urem <4 x i32> [[VAL1]], [[VAL2]]
+  ui %= ui2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = srem <2 x i64> [[VAL1]], [[VAL2]]
+  sl %= sl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = urem <2 x i64> [[VAL1]], [[VAL2]]
+  ul %= ul2;
+}
+
+void test_not (void)
+{
+// CHECK-LABEL: test_not
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL]], <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+  sc = ~sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL]], <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+  uc = ~uc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL]], <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
+  bc = ~bc2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL]], <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+  ss = ~ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL]], <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+  us = ~us2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL]], <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
+  bs = ~bs2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL]], <i32 -1, i32 -1, i32 -1, i32 -1>
+  si = ~si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL]], <i32 -1, i32 -1, i32 -1, i32 -1>
+  ui = ~ui2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL]], <i32 -1, i32 -1, i32 -1, i32 -1>
+  bi = ~bi2;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL]], <i64 -1, i64 -1>
+  sl = ~sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL]], <i64 -1, i64 -1>
+  ul = ~ul2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL]], <i64 -1, i64 -1>
+  bl = ~bl2;
+}
+
+void test_and (void)
+{
+// CHECK-LABEL: test_and
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = and <16 x i8> [[VAL2]], [[VAL1]]
+  sc = sc & sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = and <16 x i8> [[VAL2]], [[VAL1]]
+  sc = sc & bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = and <16 x i8> [[VAL2]], [[VAL1]]
+  sc = bc & sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = and <16 x i8> [[VAL2]], [[VAL1]]
+  uc = uc & uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = and <16 x i8> [[VAL2]], [[VAL1]]
+  uc = uc & bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = and <16 x i8> [[VAL2]], [[VAL1]]
+  uc = bc & uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = and <16 x i8> [[VAL2]], [[VAL1]]
+  bc = bc & bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = and <8 x i16> [[VAL2]], [[VAL1]]
+  ss = ss & ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = and <8 x i16> [[VAL2]], [[VAL1]]
+  ss = ss & bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = and <8 x i16> [[VAL2]], [[VAL1]]
+  ss = bs & ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = and <8 x i16> [[VAL2]], [[VAL1]]
+  us = us & us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = and <8 x i16> [[VAL2]], [[VAL1]]
+  us = us & bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = and <8 x i16> [[VAL2]], [[VAL1]]
+  us = bs & us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = and <8 x i16> [[VAL2]], [[VAL1]]
+  bs = bs & bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = and <4 x i32> [[VAL2]], [[VAL1]]
+  si = si & si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = and <4 x i32> [[VAL2]], [[VAL1]]
+  si = si & bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = and <4 x i32> [[VAL2]], [[VAL1]]
+  si = bi & si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = and <4 x i32> [[VAL2]], [[VAL1]]
+  ui = ui & ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = and <4 x i32> [[VAL2]], [[VAL1]]
+  ui = ui & bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = and <4 x i32> [[VAL2]], [[VAL1]]
+  ui = bi & ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = and <4 x i32> [[VAL2]], [[VAL1]]
+  bi = bi & bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = and <2 x i64> [[VAL2]], [[VAL1]]
+  sl = sl & sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = and <2 x i64> [[VAL2]], [[VAL1]]
+  sl = sl & bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = and <2 x i64> [[VAL2]], [[VAL1]]
+  sl = bl & sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = and <2 x i64> [[VAL2]], [[VAL1]]
+  ul = ul & ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = and <2 x i64> [[VAL2]], [[VAL1]]
+  ul = ul & bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = and <2 x i64> [[VAL2]], [[VAL1]]
+  ul = bl & ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = and <2 x i64> [[VAL2]], [[VAL1]]
+  bl = bl & bl2;
+}
+
+void test_and_assign (void)
+{
+// CHECK-LABEL: test_and_assign
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = and <16 x i8> [[VAL1]], [[VAL2]]
+  sc &= sc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = and <16 x i8> [[VAL1]], [[VAL2]]
+  sc &= bc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = and <16 x i8> [[VAL1]], [[VAL2]]
+  uc &= uc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = and <16 x i8> [[VAL1]], [[VAL2]]
+  uc &= bc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: %{{.*}} = and <16 x i8> [[VAL1]], [[VAL2]]
+  bc &= bc2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = and <8 x i16> [[VAL1]], [[VAL2]]
+  ss &= ss2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = and <8 x i16> [[VAL1]], [[VAL2]]
+  ss &= bs2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = and <8 x i16> [[VAL1]], [[VAL2]]
+  us &= us2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = and <8 x i16> [[VAL1]], [[VAL2]]
+  us &= bs2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: %{{.*}} = and <8 x i16> [[VAL1]], [[VAL2]]
+  bs &= bs2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = and <4 x i32> [[VAL1]], [[VAL2]]
+  si &= si2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = and <4 x i32> [[VAL1]], [[VAL2]]
+  si &= bi2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = and <4 x i32> [[VAL1]], [[VAL2]]
+  ui &= ui2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = and <4 x i32> [[VAL1]], [[VAL2]]
+  ui &= bi2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: %{{.*}} = and <4 x i32> [[VAL1]], [[VAL2]]
+  bi &= bi2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = and <2 x i64> [[VAL1]], [[VAL2]]
+  sl &= sl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = and <2 x i64> [[VAL1]], [[VAL2]]
+  sl &= bl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = and <2 x i64> [[VAL1]], [[VAL2]]
+  ul &= ul2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = and <2 x i64> [[VAL1]], [[VAL2]]
+  ul &= bl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: %{{.*}} = and <2 x i64> [[VAL1]], [[VAL2]]
+  bl &= bl2;
+}
+
+void test_or (void)
+{
+// CHECK-LABEL: test_or
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = or <16 x i8> [[VAL2]], [[VAL1]]
+  sc = sc | sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = or <16 x i8> [[VAL2]], [[VAL1]]
+  sc = sc | bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = or <16 x i8> [[VAL2]], [[VAL1]]
+  sc = bc | sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = or <16 x i8> [[VAL2]], [[VAL1]]
+  uc = uc | uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = or <16 x i8> [[VAL2]], [[VAL1]]
+  uc = uc | bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = or <16 x i8> [[VAL2]], [[VAL1]]
+  uc = bc | uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = or <16 x i8> [[VAL2]], [[VAL1]]
+  bc = bc | bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = or <8 x i16> [[VAL2]], [[VAL1]]
+  ss = ss | ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = or <8 x i16> [[VAL2]], [[VAL1]]
+  ss = ss | bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = or <8 x i16> [[VAL2]], [[VAL1]]
+  ss = bs | ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = or <8 x i16> [[VAL2]], [[VAL1]]
+  us = us | us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = or <8 x i16> [[VAL2]], [[VAL1]]
+  us = us | bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = or <8 x i16> [[VAL2]], [[VAL1]]
+  us = bs | us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = or <8 x i16> [[VAL2]], [[VAL1]]
+  bs = bs | bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = or <4 x i32> [[VAL2]], [[VAL1]]
+  si = si | si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = or <4 x i32> [[VAL2]], [[VAL1]]
+  si = si | bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = or <4 x i32> [[VAL2]], [[VAL1]]
+  si = bi | si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = or <4 x i32> [[VAL2]], [[VAL1]]
+  ui = ui | ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = or <4 x i32> [[VAL2]], [[VAL1]]
+  ui = ui | bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = or <4 x i32> [[VAL2]], [[VAL1]]
+  ui = bi | ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = or <4 x i32> [[VAL2]], [[VAL1]]
+  bi = bi | bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = or <2 x i64> [[VAL2]], [[VAL1]]
+  sl = sl | sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = or <2 x i64> [[VAL2]], [[VAL1]]
+  sl = sl | bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = or <2 x i64> [[VAL2]], [[VAL1]]
+  sl = bl | sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = or <2 x i64> [[VAL2]], [[VAL1]]
+  ul = ul | ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = or <2 x i64> [[VAL2]], [[VAL1]]
+  ul = ul | bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = or <2 x i64> [[VAL2]], [[VAL1]]
+  ul = bl | ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = or <2 x i64> [[VAL2]], [[VAL1]]
+  bl = bl | bl2;
+}
+
+void test_or_assign (void)
+{
+// CHECK-LABEL: test_or_assign
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = or <16 x i8> [[VAL1]], [[VAL2]]
+  sc |= sc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = or <16 x i8> [[VAL1]], [[VAL2]]
+  sc |= bc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = or <16 x i8> [[VAL1]], [[VAL2]]
+  uc |= uc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = or <16 x i8> [[VAL1]], [[VAL2]]
+  uc |= bc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: %{{.*}} = or <16 x i8> [[VAL1]], [[VAL2]]
+  bc |= bc2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = or <8 x i16> [[VAL1]], [[VAL2]]
+  ss |= ss2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = or <8 x i16> [[VAL1]], [[VAL2]]
+  ss |= bs2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = or <8 x i16> [[VAL1]], [[VAL2]]
+  us |= us2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = or <8 x i16> [[VAL1]], [[VAL2]]
+  us |= bs2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: %{{.*}} = or <8 x i16> [[VAL1]], [[VAL2]]
+  bs |= bs2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = or <4 x i32> [[VAL1]], [[VAL2]]
+  si |= si2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = or <4 x i32> [[VAL1]], [[VAL2]]
+  si |= bi2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = or <4 x i32> [[VAL1]], [[VAL2]]
+  ui |= ui2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = or <4 x i32> [[VAL1]], [[VAL2]]
+  ui |= bi2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: %{{.*}} = or <4 x i32> [[VAL1]], [[VAL2]]
+  bi |= bi2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = or <2 x i64> [[VAL1]], [[VAL2]]
+  sl |= sl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = or <2 x i64> [[VAL1]], [[VAL2]]
+  sl |= bl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = or <2 x i64> [[VAL1]], [[VAL2]]
+  ul |= ul2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = or <2 x i64> [[VAL1]], [[VAL2]]
+  ul |= bl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: %{{.*}} = or <2 x i64> [[VAL1]], [[VAL2]]
+  bl |= bl2;
+}
+
+void test_xor (void)
+{
+// CHECK-LABEL: test_xor
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL1]], [[VAL2]]
+  sc = sc ^ sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL1]], [[VAL2]]
+  sc = sc ^ bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL1]], [[VAL2]]
+  sc = bc ^ sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL1]], [[VAL2]]
+  uc = uc ^ uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL1]], [[VAL2]]
+  uc = uc ^ bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL1]], [[VAL2]]
+  uc = bc ^ uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL1]], [[VAL2]]
+  bc = bc ^ bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL1]], [[VAL2]]
+  ss = ss ^ ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL1]], [[VAL2]]
+  ss = ss ^ bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL1]], [[VAL2]]
+  ss = bs ^ ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL1]], [[VAL2]]
+  us = us ^ us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL1]], [[VAL2]]
+  us = us ^ bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL1]], [[VAL2]]
+  us = bs ^ us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL1]], [[VAL2]]
+  bs = bs ^ bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL1]], [[VAL2]]
+  si = si ^ si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL1]], [[VAL2]]
+  si = si ^ bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL1]], [[VAL2]]
+  si = bi ^ si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL1]], [[VAL2]]
+  ui = ui ^ ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL1]], [[VAL2]]
+  ui = ui ^ bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL1]], [[VAL2]]
+  ui = bi ^ ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL1]], [[VAL2]]
+  bi = bi ^ bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL1]], [[VAL2]]
+  sl = sl ^ sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL1]], [[VAL2]]
+  sl = sl ^ bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL1]], [[VAL2]]
+  sl = bl ^ sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL1]], [[VAL2]]
+  ul = ul ^ ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL1]], [[VAL2]]
+  ul = ul ^ bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL1]], [[VAL2]]
+  ul = bl ^ ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL1]], [[VAL2]]
+  bl = bl ^ bl2;
+}
+
+void test_xor_assign (void)
+{
+// CHECK-LABEL: test_xor_assign
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL2]], [[VAL1]]
+  sc ^= sc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL2]], [[VAL1]]
+  sc ^= bc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL2]], [[VAL1]]
+  uc ^= uc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL2]], [[VAL1]]
+  uc ^= bc2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: %{{.*}} = xor <16 x i8> [[VAL2]], [[VAL1]]
+  bc ^= bc2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL2]], [[VAL1]]
+  ss ^= ss2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL2]], [[VAL1]]
+  ss ^= bs2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL2]], [[VAL1]]
+  us ^= us2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL2]], [[VAL1]]
+  us ^= bs2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: %{{.*}} = xor <8 x i16> [[VAL2]], [[VAL1]]
+  bs ^= bs2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL2]], [[VAL1]]
+  si ^= si2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL2]], [[VAL1]]
+  si ^= bi2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL2]], [[VAL1]]
+  ui ^= ui2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL2]], [[VAL1]]
+  ui ^= bi2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: %{{.*}} = xor <4 x i32> [[VAL2]], [[VAL1]]
+  bi ^= bi2;
+
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL2]], [[VAL1]]
+  sl ^= sl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL2]], [[VAL1]]
+  sl ^= bl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL2]], [[VAL1]]
+  ul ^= ul2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL2]], [[VAL1]]
+  ul ^= bl2;
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: %{{.*}} = xor <2 x i64> [[VAL2]], [[VAL1]]
+  bl ^= bl2;
+}
+
+void test_sl (void)
+{
+// CHECK-LABEL: test_sl
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  sc = sc << sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  sc = sc << uc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <16 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <16 x i32> [[T2]], <16 x i32> undef, <16 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = trunc <16 x i32> [[T3]] to <16 x i8>
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  sc = sc << cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
+  sc = sc << 5;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  uc = uc << sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  uc = uc << uc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <16 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <16 x i32> [[T2]], <16 x i32> undef, <16 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = trunc <16 x i32> [[T3]] to <16 x i8>
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  uc = uc << cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
+  uc = uc << 5;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  ss = ss << ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  ss = ss << us2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <8 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <8 x i32> [[T2]], <8 x i32> undef, <8 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = trunc <8 x i32> [[T3]] to <8 x i16>
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  ss = ss << cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
+  ss = ss << 5;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  us = us << ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  us = us << us2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <8 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <8 x i32> [[T2]], <8 x i32> undef, <8 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = trunc <8 x i32> [[T3]] to <8 x i16>
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  us = us << cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
+  us = us << 5;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  si = si << si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  si = si << ui2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T3:%[^ ]+]] = insertelement <4 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[CNT:%[^ ]+]] = shufflevector <4 x i32> [[T3]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  si = si << cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], <i32 5, i32 5, i32 5, i32 5>
+  si = si << 5;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  ui = ui << si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  ui = ui << ui2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T3:%[^ ]+]] = insertelement <4 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[CNT:%[^ ]+]] = shufflevector <4 x i32> [[T3]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  ui = ui << cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], <i32 5, i32 5, i32 5, i32 5>
+  ui = ui << 5;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  sl = sl << sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  sl = sl << ul2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <2 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <2 x i32> [[T2]], <2 x i32> undef, <2 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = zext <2 x i32> [[T3]] to <2 x i64>
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  sl = sl << cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], <i64 5, i64 5>
+  sl = sl << 5;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  ul = ul << sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  ul = ul << ul2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <2 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <2 x i32> [[T2]], <2 x i32> undef, <2 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = zext <2 x i32> [[T3]] to <2 x i64>
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  ul = ul << cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], <i64 5, i64 5>
+  ul = ul << 5;
+}
+
+void test_sl_assign (void)
+{
+// CHECK-LABEL: test_sl_assign
+
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  sc <<= sc2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  sc <<= uc2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <16 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <16 x i32> [[T2]], <16 x i32> undef, <16 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[CNT:%[^ ]+]] = trunc <16 x i32> [[T3]] to <16 x i8>
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  sc <<= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
+  sc <<= 5;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  uc <<= sc2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  uc <<= uc2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <16 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <16 x i32> [[T2]], <16 x i32> undef, <16 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[CNT:%[^ ]+]] = trunc <16 x i32> [[T3]] to <16 x i8>
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], [[CNT]]
+  uc <<= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = shl <16 x i8> [[VAL]], <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
+  uc <<= 5;
+
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  ss <<= ss2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  ss <<= us2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <8 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <8 x i32> [[T2]], <8 x i32> undef, <8 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[CNT:%[^ ]+]] = trunc <8 x i32> [[T3]] to <8 x i16>
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  ss <<= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
+  ss <<= 5;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  us <<= ss2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  us <<= us2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <8 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <8 x i32> [[T2]], <8 x i32> undef, <8 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[CNT:%[^ ]+]] = trunc <8 x i32> [[T3]] to <8 x i16>
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], [[CNT]]
+  us <<= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = shl <8 x i16> [[VAL]], <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
+  us <<= 5;
+
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  si <<= si2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  si <<= ui2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T3:%[^ ]+]] = insertelement <4 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[CNT:%[^ ]+]] = shufflevector <4 x i32> [[T3]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  si <<= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], <i32 5, i32 5, i32 5, i32 5>
+  si <<= 5;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  ui <<= si2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  ui <<= ui2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T3:%[^ ]+]] = insertelement <4 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[CNT:%[^ ]+]] = shufflevector <4 x i32> [[T3]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], [[CNT]]
+  ui <<= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = shl <4 x i32> [[VAL]], <i32 5, i32 5, i32 5, i32 5>
+  ui <<= 5;
+
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  sl <<= sl2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  sl <<= ul2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <2 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <2 x i32> [[T2]], <2 x i32> undef, <2 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[CNT:%[^ ]+]] = zext <2 x i32> [[T3]] to <2 x i64>
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  sl <<= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], <i64 5, i64 5>
+  sl <<= 5;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  ul <<= sl2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  ul <<= ul2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <2 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <2 x i32> [[T2]], <2 x i32> undef, <2 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[CNT:%[^ ]+]] = zext <2 x i32> [[T3]] to <2 x i64>
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], [[CNT]]
+  ul <<= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = shl <2 x i64> [[VAL]], <i64 5, i64 5>
+  ul <<= 5;
+}
+
+void test_sr (void)
+{
+// CHECK-LABEL: test_sr
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = ashr <16 x i8> [[VAL]], [[CNT]]
+  sc = sc >> sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = ashr <16 x i8> [[VAL]], [[CNT]]
+  sc = sc >> uc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <16 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <16 x i32> [[T2]], <16 x i32> undef, <16 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = trunc <16 x i32> [[T3]] to <16 x i8>
+// CHECK: %{{.*}} = ashr <16 x i8> [[VAL]], [[CNT]]
+  sc = sc >> cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = ashr <16 x i8> [[VAL]], <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
+  sc = sc >> 5;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: %{{.*}} = lshr <16 x i8> [[VAL]], [[CNT]]
+  uc = uc >> sc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: %{{.*}} = lshr <16 x i8> [[VAL]], [[CNT]]
+  uc = uc >> uc2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <16 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <16 x i32> [[T2]], <16 x i32> undef, <16 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = trunc <16 x i32> [[T3]] to <16 x i8>
+// CHECK: %{{.*}} = lshr <16 x i8> [[VAL]], [[CNT]]
+  uc = uc >> cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = lshr <16 x i8> [[VAL]], <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
+  uc = uc >> 5;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = ashr <8 x i16> [[VAL]], [[CNT]]
+  ss = ss >> ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = ashr <8 x i16> [[VAL]], [[CNT]]
+  ss = ss >> us2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <8 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <8 x i32> [[T2]], <8 x i32> undef, <8 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = trunc <8 x i32> [[T3]] to <8 x i16>
+// CHECK: %{{.*}} = ashr <8 x i16> [[VAL]], [[CNT]]
+  ss = ss >> cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = ashr <8 x i16> [[VAL]], <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
+  ss = ss >> 5;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: %{{.*}} = lshr <8 x i16> [[VAL]], [[CNT]]
+  us = us >> ss2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: %{{.*}} = lshr <8 x i16> [[VAL]], [[CNT]]
+  us = us >> us2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <8 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <8 x i32> [[T2]], <8 x i32> undef, <8 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = trunc <8 x i32> [[T3]] to <8 x i16>
+// CHECK: %{{.*}} = lshr <8 x i16> [[VAL]], [[CNT]]
+  us = us >> cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = lshr <8 x i16> [[VAL]], <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
+  us = us >> 5;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = ashr <4 x i32> [[VAL]], [[CNT]]
+  si = si >> si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = ashr <4 x i32> [[VAL]], [[CNT]]
+  si = si >> ui2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T3:%[^ ]+]] = insertelement <4 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[CNT:%[^ ]+]] = shufflevector <4 x i32> [[T3]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK: %{{.*}} = ashr <4 x i32> [[VAL]], [[CNT]]
+  si = si >> cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = ashr <4 x i32> [[VAL]], <i32 5, i32 5, i32 5, i32 5>
+  si = si >> 5;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: %{{.*}} = lshr <4 x i32> [[VAL]], [[CNT]]
+  ui = ui >> si2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: %{{.*}} = lshr <4 x i32> [[VAL]], [[CNT]]
+  ui = ui >> ui2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T3:%[^ ]+]] = insertelement <4 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[CNT:%[^ ]+]] = shufflevector <4 x i32> [[T3]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK: %{{.*}} = lshr <4 x i32> [[VAL]], [[CNT]]
+  ui = ui >> cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = lshr <4 x i32> [[VAL]], <i32 5, i32 5, i32 5, i32 5>
+  ui = ui >> 5;
+
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = ashr <2 x i64> [[VAL]], [[CNT]]
+  sl = sl >> sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = ashr <2 x i64> [[VAL]], [[CNT]]
+  sl = sl >> ul2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <2 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <2 x i32> [[T2]], <2 x i32> undef, <2 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = zext <2 x i32> [[T3]] to <2 x i64>
+// CHECK: %{{.*}} = ashr <2 x i64> [[VAL]], [[CNT]]
+  sl = sl >> cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = ashr <2 x i64> [[VAL]], <i64 5, i64 5>
+  sl = sl >> 5;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: %{{.*}} = lshr <2 x i64> [[VAL]], [[CNT]]
+  ul = ul >> sl2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: %{{.*}} = lshr <2 x i64> [[VAL]], [[CNT]]
+  ul = ul >> ul2;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <2 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <2 x i32> [[T2]], <2 x i32> undef, <2 x i32> zeroinitializer
+// CHECK: [[CNT:%[^ ]+]] = zext <2 x i32> [[T3]] to <2 x i64>
+// CHECK: %{{.*}} = lshr <2 x i64> [[VAL]], [[CNT]]
+  ul = ul >> cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = lshr <2 x i64> [[VAL]], <i64 5, i64 5>
+  ul = ul >> 5;
+}
+
+void test_sr_assign (void)
+{
+// CHECK-LABEL: test_sr_assign
+
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = ashr <16 x i8> [[VAL]], [[CNT]]
+  sc >>= sc2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = ashr <16 x i8> [[VAL]], [[CNT]]
+  sc >>= uc2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <16 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <16 x i32> [[T2]], <16 x i32> undef, <16 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[CNT:%[^ ]+]] = trunc <16 x i32> [[T3]] to <16 x i8>
+// CHECK: %{{.*}} = ashr <16 x i8> [[VAL]], [[CNT]]
+  sc >>= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: %{{.*}} = ashr <16 x i8> [[VAL]], <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
+  sc >>= 5;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = lshr <16 x i8> [[VAL]], [[CNT]]
+  uc >>= sc2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = lshr <16 x i8> [[VAL]], [[CNT]]
+  uc >>= uc2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <16 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <16 x i32> [[T2]], <16 x i32> undef, <16 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[CNT:%[^ ]+]] = trunc <16 x i32> [[T3]] to <16 x i8>
+// CHECK: %{{.*}} = lshr <16 x i8> [[VAL]], [[CNT]]
+  uc >>= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: %{{.*}} = lshr <16 x i8> [[VAL]], <i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5, i8 5>
+  uc >>= 5;
+
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = ashr <8 x i16> [[VAL]], [[CNT]]
+  ss >>= ss2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = ashr <8 x i16> [[VAL]], [[CNT]]
+  ss >>= us2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <8 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <8 x i32> [[T2]], <8 x i32> undef, <8 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[CNT:%[^ ]+]] = trunc <8 x i32> [[T3]] to <8 x i16>
+// CHECK: %{{.*}} = ashr <8 x i16> [[VAL]], [[CNT]]
+  ss >>= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: %{{.*}} = ashr <8 x i16> [[VAL]], <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
+  ss >>= 5;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = lshr <8 x i16> [[VAL]], [[CNT]]
+  us >>= ss2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = lshr <8 x i16> [[VAL]], [[CNT]]
+  us >>= us2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <8 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <8 x i32> [[T2]], <8 x i32> undef, <8 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[CNT:%[^ ]+]] = trunc <8 x i32> [[T3]] to <8 x i16>
+// CHECK: %{{.*}} = lshr <8 x i16> [[VAL]], [[CNT]]
+  us >>= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: %{{.*}} = lshr <8 x i16> [[VAL]], <i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5, i16 5>
+  us >>= 5;
+
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = ashr <4 x i32> [[VAL]], [[CNT]]
+  si >>= si2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = ashr <4 x i32> [[VAL]], [[CNT]]
+  si >>= ui2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T3:%[^ ]+]] = insertelement <4 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[CNT:%[^ ]+]] = shufflevector <4 x i32> [[T3]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = ashr <4 x i32> [[VAL]], [[CNT]]
+  si >>= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: %{{.*}} = ashr <4 x i32> [[VAL]], <i32 5, i32 5, i32 5, i32 5>
+  si >>= 5;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = lshr <4 x i32> [[VAL]], [[CNT]]
+  ui >>= si2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = lshr <4 x i32> [[VAL]], [[CNT]]
+  ui >>= ui2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T3:%[^ ]+]] = insertelement <4 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[CNT:%[^ ]+]] = shufflevector <4 x i32> [[T3]], <4 x i32> undef, <4 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = lshr <4 x i32> [[VAL]], [[CNT]]
+  ui >>= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: %{{.*}} = lshr <4 x i32> [[VAL]], <i32 5, i32 5, i32 5, i32 5>
+  ui >>= 5;
+
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = ashr <2 x i64> [[VAL]], [[CNT]]
+  sl >>= sl2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = ashr <2 x i64> [[VAL]], [[CNT]]
+  sl >>= ul2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <2 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <2 x i32> [[T2]], <2 x i32> undef, <2 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[CNT:%[^ ]+]] = zext <2 x i32> [[T3]] to <2 x i64>
+// CHECK: %{{.*}} = ashr <2 x i64> [[VAL]], [[CNT]]
+  sl >>= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: %{{.*}} = ashr <2 x i64> [[VAL]], <i64 5, i64 5>
+  sl >>= 5;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = lshr <2 x i64> [[VAL]], [[CNT]]
+  ul >>= sl2;
+// CHECK: [[CNT:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = lshr <2 x i64> [[VAL]], [[CNT]]
+  ul >>= ul2;
+// CHECK: [[T1:%[^ ]+]] = load volatile i32, i32* @cnt
+// CHECK: [[T2:%[^ ]+]] = insertelement <2 x i32> undef, i32 [[T1]], i32 0
+// CHECK: [[T3:%[^ ]+]] = shufflevector <2 x i32> [[T2]], <2 x i32> undef, <2 x i32> zeroinitializer
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[CNT:%[^ ]+]] = zext <2 x i32> [[T3]] to <2 x i64>
+// CHECK: %{{.*}} = lshr <2 x i64> [[VAL]], [[CNT]]
+  ul >>= cnt;
+// CHECK: [[VAL:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: %{{.*}} = lshr <2 x i64> [[VAL]], <i64 5, i64 5>
+  ul >>= 5;
+}
+
+
+void test_cmpeq (void)
+{
+// CHECK-LABEL: test_cmpeq
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = sc == sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = sc == bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc == sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = uc == uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = uc == bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc == uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc == bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = ss == ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = ss == bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs == ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = us == us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = us == bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs == us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs == bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = si == si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = si == bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi == si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = ui == ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = ui == bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi == ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi == bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = sl == sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = sl == bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl == sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = ul == ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = ul == bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl == ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp eq <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl == bl2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[CMP:%[^ ]+]] = fcmp oeq <2 x double> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = fd == fd2;
+}
+
+void test_cmpne (void)
+{
+// CHECK-LABEL: test_cmpne
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = sc != sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = sc != bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc != sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = uc != uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = uc != bc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc != uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc != bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = ss != ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = ss != bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs != ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = us != us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = us != bs2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs != us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs != bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = si != si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = si != bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi != si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = ui != ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = ui != bi2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi != ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi != bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = sl != sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = sl != bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl != sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = ul != ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = ul != bl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl != ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp ne <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl != bl2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[CMP:%[^ ]+]] = fcmp une <2 x double> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = fd != fd2;
+}
+
+void test_cmpge (void)
+{
+// CHECK-LABEL: test_cmpge
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[CMP:%[^ ]+]] = icmp sge <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = sc >= sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[CMP:%[^ ]+]] = icmp uge <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = uc >= uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp uge <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc >= bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[CMP:%[^ ]+]] = icmp sge <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = ss >= ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[CMP:%[^ ]+]] = icmp uge <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = us >= us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp uge <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs >= bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[CMP:%[^ ]+]] = icmp sge <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = si >= si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[CMP:%[^ ]+]] = icmp uge <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = ui >= ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp uge <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi >= bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[CMP:%[^ ]+]] = icmp sge <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = sl >= sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[CMP:%[^ ]+]] = icmp uge <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = ul >= ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp uge <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl >= bl2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[CMP:%[^ ]+]] = fcmp oge <2 x double> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = fd >= fd2;
+}
+
+void test_cmpgt (void)
+{
+// CHECK-LABEL: test_cmpgt
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[CMP:%[^ ]+]] = icmp sgt <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = sc > sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ugt <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = uc > uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ugt <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc > bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[CMP:%[^ ]+]] = icmp sgt <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = ss > ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[CMP:%[^ ]+]] = icmp ugt <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = us > us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp ugt <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs > bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[CMP:%[^ ]+]] = icmp sgt <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = si > si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[CMP:%[^ ]+]] = icmp ugt <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = ui > ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp ugt <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi > bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[CMP:%[^ ]+]] = icmp sgt <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = sl > sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[CMP:%[^ ]+]] = icmp ugt <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = ul > ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp ugt <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl > bl2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[CMP:%[^ ]+]] = fcmp ogt <2 x double> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = fd > fd2;
+}
+
+void test_cmple (void)
+{
+// CHECK-LABEL: test_cmple
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[CMP:%[^ ]+]] = icmp sle <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = sc <= sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ule <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = uc <= uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ule <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc <= bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[CMP:%[^ ]+]] = icmp sle <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = ss <= ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[CMP:%[^ ]+]] = icmp ule <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = us <= us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp ule <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs <= bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[CMP:%[^ ]+]] = icmp sle <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = si <= si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[CMP:%[^ ]+]] = icmp ule <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = ui <= ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp ule <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi <= bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[CMP:%[^ ]+]] = icmp sle <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = sl <= sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[CMP:%[^ ]+]] = icmp ule <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = ul <= ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp ule <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl <= bl2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[CMP:%[^ ]+]] = fcmp ole <2 x double> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = fd <= fd2;
+}
+
+void test_cmplt (void)
+{
+// CHECK-LABEL: test_cmplt
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @sc2
+// CHECK: [[CMP:%[^ ]+]] = icmp slt <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = sc < sc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @uc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ult <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = uc < uc2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <16 x i8>, <16 x i8>* @bc2
+// CHECK: [[CMP:%[^ ]+]] = icmp ult <16 x i8> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <16 x i1> [[CMP]] to <16 x i8>
+  bc = bc < bc2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @ss2
+// CHECK: [[CMP:%[^ ]+]] = icmp slt <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = ss < ss2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @us2
+// CHECK: [[CMP:%[^ ]+]] = icmp ult <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = us < us2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <8 x i16>, <8 x i16>* @bs2
+// CHECK: [[CMP:%[^ ]+]] = icmp ult <8 x i16> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <8 x i1> [[CMP]] to <8 x i16>
+  bs = bs < bs2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @si2
+// CHECK: [[CMP:%[^ ]+]] = icmp slt <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = si < si2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @ui2
+// CHECK: [[CMP:%[^ ]+]] = icmp ult <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = ui < ui2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <4 x i32>, <4 x i32>* @bi2
+// CHECK: [[CMP:%[^ ]+]] = icmp ult <4 x i32> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <4 x i1> [[CMP]] to <4 x i32>
+  bi = bi < bi2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @sl2
+// CHECK: [[CMP:%[^ ]+]] = icmp slt <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = sl < sl2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @ul2
+// CHECK: [[CMP:%[^ ]+]] = icmp ult <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = ul < ul2;
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x i64>, <2 x i64>* @bl2
+// CHECK: [[CMP:%[^ ]+]] = icmp ult <2 x i64> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = bl < bl2;
+
+// CHECK: [[VAL1:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd
+// CHECK: [[VAL2:%[^ ]+]] = load volatile <2 x double>, <2 x double>* @fd2
+// CHECK: [[CMP:%[^ ]+]] = fcmp olt <2 x double> [[VAL1]], [[VAL2]]
+// CHECK: %{{.*}} = sext <2 x i1> [[CMP]] to <2 x i64>
+  bl = fd < fd2;
+}
+
diff --git a/clang/test/Preprocessor/predefined-arch-macros.c b/clang/test/Preprocessor/predefined-arch-macros.c
index 5126ef3..14fc494 100644
--- a/clang/test/Preprocessor/predefined-arch-macros.c
+++ b/clang/test/Preprocessor/predefined-arch-macros.c
@@ -1748,3 +1748,12 @@
 // RUN:   | FileCheck %s -check-prefix=CHECK_SYSTEMZ_HTM
 //
 // CHECK_SYSTEMZ_HTM: #define __HTM__ 1
+//
+// RUN: %clang -fzvector -E -dM %s -o - 2>&1 \
+// RUN:     -target s390x-unknown-linux \
+// RUN:   | FileCheck %s -check-prefix=CHECK_SYSTEMZ_ZVECTOR
+// RUN: %clang -mzvector -E -dM %s -o - 2>&1 \
+// RUN:     -target s390x-unknown-linux \
+// RUN:   | FileCheck %s -check-prefix=CHECK_SYSTEMZ_ZVECTOR
+//
+// CHECK_SYSTEMZ_ZVECTOR: #define __VEC__ 10301
diff --git a/clang/test/Sema/zvector.c b/clang/test/Sema/zvector.c
new file mode 100644
index 0000000..5220a62
--- /dev/null
+++ b/clang/test/Sema/zvector.c
@@ -0,0 +1,1009 @@
+// RUN: %clang_cc1 -triple s390x-linux-gnu -fzvector \
+// RUN:  -fno-lax-vector-conversions -W -Wall -Wconversion \
+// RUN:  -Werror -fsyntax-only -verify %s
+
+vector signed char sc, sc2;
+vector unsigned char uc, uc2;
+vector bool char bc, bc2;
+
+vector signed short ss, ss2;
+vector unsigned short us, us2;
+vector bool short bs, bs2;
+
+vector signed int si, si2;
+vector unsigned int ui, ui2;
+vector bool int bi, bi2;
+
+vector signed long long sl, sl2;
+vector unsigned long long ul, ul2;
+vector bool long long bl, bl2;
+
+vector double fd, fd2;
+
+vector long ll; // expected-error {{cannot use 'long' with '__vector'}}
+vector float ff; // expected-error {{cannot use 'float' with '__vector'}}
+
+signed char sc_scalar;
+unsigned char uc_scalar;
+
+signed short ss_scalar;
+unsigned short us_scalar;
+
+signed int si_scalar;
+unsigned int ui_scalar;
+
+signed long sl_scalar;
+unsigned long ul_scalar;
+
+double fd_scalar;
+
+void foo(void)
+{
+  // -------------------------------------------------------------------------
+  // Test assignment.
+  // -------------------------------------------------------------------------
+
+  sc = sc2;
+  uc = uc2;
+  bc = bc2;
+
+  ss = ss2;
+  us = us2;
+  bs = bs2;
+
+  si = si2;
+  ui = ui2;
+  bi = bi2;
+
+  sl = sl2;
+  ul = ul2;
+  bl = bl2;
+  fd = fd2;
+
+  sc = uc2; // expected-error {{incompatible type}}
+  sc = bc2; // expected-error {{incompatible type}}
+  uc = sc2; // expected-error {{incompatible type}}
+  uc = bc2; // expected-error {{incompatible type}}
+  bc = sc2; // expected-error {{incompatible type}}
+  bc = uc2; // expected-error {{incompatible type}}
+
+  sc = sc_scalar; // expected-error {{incompatible type}}
+  sc = uc_scalar; // expected-error {{incompatible type}}
+  uc = sc_scalar; // expected-error {{incompatible type}}
+  uc = uc_scalar; // expected-error {{incompatible type}}
+  bc = sc_scalar; // expected-error {{incompatible type}}
+  bc = uc_scalar; // expected-error {{incompatible type}}
+
+  sc = ss2; // expected-error {{incompatible type}}
+  sc = si2; // expected-error {{incompatible type}}
+  sc = sl2; // expected-error {{incompatible type}}
+  sc = fd2; // expected-error {{incompatible type}}
+
+  ss = sc2; // expected-error {{incompatible type}}
+  si = sc2; // expected-error {{incompatible type}}
+  sl = sc2; // expected-error {{incompatible type}}
+  fd = sc2; // expected-error {{incompatible type}}
+
+  uc = us2; // expected-error {{incompatible type}}
+  uc = ui2; // expected-error {{incompatible type}}
+  uc = ul2; // expected-error {{incompatible type}}
+  uc = fd2; // expected-error {{incompatible type}}
+
+  us = uc2; // expected-error {{incompatible type}}
+  ui = uc2; // expected-error {{incompatible type}}
+  ul = uc2; // expected-error {{incompatible type}}
+  fd = uc2; // expected-error {{incompatible type}}
+
+  bc = us2; // expected-error {{incompatible type}}
+  bc = ui2; // expected-error {{incompatible type}}
+  bc = ul2; // expected-error {{incompatible type}}
+  bc = fd2; // expected-error {{incompatible type}}
+
+  bs = bc2; // expected-error {{incompatible type}}
+  bi = bc2; // expected-error {{incompatible type}}
+  bl = bc2; // expected-error {{incompatible type}}
+  fd = bc2; // expected-error {{incompatible type}}
+
+  // -------------------------------------------------------------------------
+  // Test casts to same element width.
+  // -------------------------------------------------------------------------
+
+  sc = (vector signed char)bc2;
+  bc = (vector bool char)uc2;
+  uc = (vector unsigned char)sc2;
+
+  ss = (vector signed short)bs2;
+  bs = (vector bool short)us2;
+  us = (vector unsigned short)ss2;
+
+  si = (vector signed int)bi2;
+  bi = (vector bool int)ui2;
+  ui = (vector unsigned int)si2;
+
+  sl = (vector signed long long)bl2;
+  bl = (vector bool long long)ul2;
+  ul = (vector unsigned long long)fd2;
+  fd = (vector double)sl2;
+
+  // -------------------------------------------------------------------------
+  // Test casts to different element width.
+  // -------------------------------------------------------------------------
+
+  sc = (vector signed char)bs2;
+  bc = (vector bool char)us2;
+  uc = (vector unsigned char)fd2;
+
+  ss = (vector signed short)bi2;
+  bs = (vector bool short)ui2;
+  us = (vector unsigned short)fd2;
+
+  si = (vector signed int)bl2;
+  bi = (vector bool int)ul2;
+  ui = (vector unsigned int)fd2;
+
+  sl = (vector signed long long)bc2;
+  bl = (vector bool long long)uc2;
+  ul = (vector unsigned long long)sc2;
+  fd = (vector double)sc2;
+
+  // -------------------------------------------------------------------------
+  // Test ++.
+  // -------------------------------------------------------------------------
+
+  ++sc2;
+  ++uc2;
+  ++bc2; // expected-error {{cannot increment}}
+
+  ++ss2;
+  ++us2;
+  ++bs2; // expected-error {{cannot increment}}
+
+  ++si2;
+  ++ui2;
+  ++bi2; // expected-error {{cannot increment}}
+
+  ++sl2;
+  ++ul2;
+  ++bl2; // expected-error {{cannot increment}}
+
+  ++fd2;
+
+  sc++;
+  uc++;
+  bc++; // expected-error {{cannot increment}}
+
+  ss++;
+  us++;
+  bs++; // expected-error {{cannot increment}}
+
+  si++;
+  ui++;
+  bi++; // expected-error {{cannot increment}}
+
+  sl++;
+  ul++;
+  bl++; // expected-error {{cannot increment}}
+
+  fd++;
+
+  // -------------------------------------------------------------------------
+  // Test --.
+  // -------------------------------------------------------------------------
+
+  --sc2;
+  --uc2;
+  --bc2; // expected-error {{cannot decrement}}
+
+  --ss2;
+  --us2;
+  --bs2; // expected-error {{cannot decrement}}
+
+  --si2;
+  --ui2;
+  --bi2; // expected-error {{cannot decrement}}
+
+  --sl2;
+  --ul2;
+  --bl2; // expected-error {{cannot decrement}}
+
+  --fd2;
+
+  sc--;
+  uc--;
+  bc--; // expected-error {{cannot decrement}}
+
+  ss--;
+  us--;
+  bs--; // expected-error {{cannot decrement}}
+
+  si--;
+  ui--;
+  bi--; // expected-error {{cannot decrement}}
+
+  sl--;
+  ul--;
+  bl--; // expected-error {{cannot decrement}}
+
+  fd--;
+
+  // -------------------------------------------------------------------------
+  // Test unary +.
+  // -------------------------------------------------------------------------
+
+  sc = +sc2;
+  uc = +uc2;
+  bc = +bc2; // expected-error {{invalid argument type}}
+
+  ss = +ss2;
+  us = +us2;
+  bs = +bs2; // expected-error {{invalid argument type}}
+
+  si = +si2;
+  ui = +ui2;
+  bi = +bi2; // expected-error {{invalid argument type}}
+
+  sl = +sl2;
+  ul = +ul2;
+  bl = +bl2; // expected-error {{invalid argument type}}
+
+  fd = +fd2;
+
+  sc = +si2; // expected-error {{assigning to}}
+  ui = +si2; // expected-error {{assigning to}}
+  ui = +bi2; // expected-error {{invalid argument type}}
+
+  // -------------------------------------------------------------------------
+  // Test unary -.
+  // -------------------------------------------------------------------------
+
+  sc = -sc2;
+  uc = -uc2;
+  bc = -bc2; // expected-error {{invalid argument type}}
+
+  ss = -ss2;
+  us = -us2;
+  bs = -bs2; // expected-error {{invalid argument type}}
+
+  si = -si2;
+  ui = -ui2;
+  bi = -bi2; // expected-error {{invalid argument type}}
+
+  sl = -sl2;
+  ul = -ul2;
+  bl = -bl2; // expected-error {{invalid argument type}}
+
+  fd = -fd2;
+
+  sc = -si2; // expected-error {{assigning to}}
+  ui = -si2; // expected-error {{assigning to}}
+  ui = -bi2; // expected-error {{invalid argument type}}
+
+  // -------------------------------------------------------------------------
+  // Test ~.
+  // -------------------------------------------------------------------------
+
+  sc = ~sc2;
+  uc = ~uc2;
+  bc = ~bc2;
+
+  ss = ~ss2;
+  us = ~us2;
+  bs = ~bs2;
+
+  si = ~si2;
+  ui = ~ui2;
+  bi = ~bi2;
+
+  sl = ~sl2;
+  ul = ~ul2;
+  bl = ~bl2;
+
+  fd = ~fd2; // expected-error {{invalid argument}}
+
+  sc = ~si2; // expected-error {{assigning to}}
+  ui = ~si2; // expected-error {{assigning to}}
+  ui = ~bi2; // expected-error {{assigning to}}
+
+  // -------------------------------------------------------------------------
+  // Test binary +.
+  // -------------------------------------------------------------------------
+
+  sc = sc + sc2;
+  sc = sc + uc2; // expected-error {{can't convert}}
+  sc = uc + sc2; // expected-error {{can't convert}}
+  sc = sc + bc2;
+  sc = bc + sc2;
+
+  uc = uc + uc2;
+  uc = sc + uc2; // expected-error {{can't convert}}
+  uc = uc + sc2; // expected-error {{can't convert}}
+  uc = bc + uc2;
+  uc = uc + bc2;
+
+  bc = bc + bc2; // expected-error {{invalid operands}}
+  bc = bc + uc2; // expected-error {{incompatible type}}
+  bc = uc + bc2; // expected-error {{incompatible type}}
+  bc = bc + sc2; // expected-error {{incompatible type}}
+  bc = sc + bc2; // expected-error {{incompatible type}}
+
+  sc = sc + sc_scalar; // expected-error {{can't convert}}
+  sc = sc + uc_scalar; // expected-error {{can't convert}}
+  sc = sc_scalar + sc; // expected-error {{can't convert}}
+  sc = uc_scalar + sc; // expected-error {{can't convert}}
+  uc = uc + sc_scalar; // expected-error {{can't convert}}
+  uc = uc + uc_scalar; // expected-error {{can't convert}}
+  uc = sc_scalar + uc; // expected-error {{can't convert}}
+  uc = uc_scalar + uc; // expected-error {{can't convert}}
+
+  ss = ss + ss2;
+  us = us + us2;
+  bs = bs + bs2; // expected-error {{invalid operands}}
+
+  si = si + si2;
+  ui = ui + ui2;
+  bi = bi + bi2; // expected-error {{invalid operands}}
+
+  sl = sl + sl2;
+  ul = ul + ul2;
+  bl = bl + bl2; // expected-error {{invalid operands}}
+
+  fd = fd + fd2;
+  fd = fd + ul2; // expected-error {{can't convert}}
+  fd = sl + fd2; // expected-error {{can't convert}}
+
+  sc += sc2;
+  sc += uc2; // expected-error {{can't convert}}
+  sc += bc2;
+
+  uc += uc2;
+  uc += sc2; // expected-error {{can't convert}}
+  uc += bc2;
+
+  bc += bc2; // expected-error {{invalid operands}}
+  bc += sc2; // expected-error {{can't convert}}
+  bc += uc2; // expected-error {{can't convert}}
+
+  sc += ss2; // expected-error {{can't convert}}
+  sc += si2; // expected-error {{can't convert}}
+  sc += sl2; // expected-error {{can't convert}}
+  sc += fd2; // expected-error {{can't convert}}
+
+  sc += sc_scalar; // expected-error {{can't convert}}
+  sc += uc_scalar; // expected-error {{can't convert}}
+  uc += sc_scalar; // expected-error {{can't convert}}
+  uc += uc_scalar; // expected-error {{can't convert}}
+
+  ss += ss2;
+  us += us2;
+  bs += bs2; // expected-error {{invalid operands}}
+
+  si += si2;
+  ui += ui2;
+  bi += bi2; // expected-error {{invalid operands}}
+
+  sl += sl2;
+  ul += ul2;
+  bl += bl2; // expected-error {{invalid operands}}
+
+  fd += fd2;
+
+  // -------------------------------------------------------------------------
+  // Test that binary + rules apply to binary - too.
+  // -------------------------------------------------------------------------
+
+  sc = sc - sc2;
+  uc = uc - uc2;
+  bc = bc - bc2; // expected-error {{invalid operands}}
+
+  sc = uc - sc2; // expected-error {{can't convert}}
+  sc = sc - bc2;
+  uc = bc - uc2;
+
+  sc -= sc2;
+  uc -= uc2;
+  bc -= bc2; // expected-error {{invalid operands}}
+
+  sc -= uc2; // expected-error {{can't convert}}
+  uc -= bc2;
+  bc -= sc2; // expected-error {{can't convert}}
+
+  ss -= ss2;
+  us -= us2;
+  bs -= bs2; // expected-error {{invalid operands}}
+
+  si -= si2;
+  ui -= ui2;
+  bi -= bi2; // expected-error {{invalid operands}}
+
+  sl -= sl2;
+  ul -= ul2;
+  bl -= bl2; // expected-error {{invalid operands}}
+
+  fd -= fd2;
+
+  // -------------------------------------------------------------------------
+  // Test that binary + rules apply to * too.  64-bit integer multiplication
+  // is not required by the spec and so isn't tested here.
+  // -------------------------------------------------------------------------
+
+  sc = sc * sc2;
+  uc = uc * uc2;
+  bc = bc * bc2; // expected-error {{invalid operands}}
+
+  sc = uc * sc2; // expected-error {{can't convert}}
+  sc = sc * bc2; // expected-error {{can't convert}}
+  uc = bc * uc2; // expected-error {{can't convert}}
+
+  sc *= sc2;
+  uc *= uc2;
+  bc *= bc2; // expected-error {{invalid operands}}
+
+  sc *= uc2; // expected-error {{can't convert}}
+  uc *= bc2; // expected-error {{can't convert}}
+  bc *= sc2; // expected-error {{can't convert}}
+
+  ss *= ss2;
+  us *= us2;
+  bs *= bs2; // expected-error {{invalid operands}}
+
+  si *= si2;
+  ui *= ui2;
+  bi *= bi2; // expected-error {{invalid operands}}
+
+  sl *= sl2;
+  ul *= ul2;
+  bl *= bl2; // expected-error {{invalid operands}}
+
+  fd *= fd2;
+
+  // -------------------------------------------------------------------------
+  // Test that * rules apply to / too.
+  // -------------------------------------------------------------------------
+
+  sc = sc / sc2;
+  uc = uc / uc2;
+  bc = bc / bc2; // expected-error {{invalid operands}}
+
+  sc = uc / sc2; // expected-error {{can't convert}}
+  sc = sc / bc2; // expected-error {{can't convert}}
+  uc = bc / uc2; // expected-error {{can't convert}}
+
+  sc /= sc2;
+  uc /= uc2;
+  bc /= bc2; // expected-error {{invalid operands}}
+
+  sc /= uc2; // expected-error {{can't convert}}
+  uc /= bc2; // expected-error {{can't convert}}
+  bc /= sc2; // expected-error {{can't convert}}
+
+  ss /= ss2;
+  us /= us2;
+  bs /= bs2; // expected-error {{invalid operands}}
+
+  si /= si2;
+  ui /= ui2;
+  bi /= bi2; // expected-error {{invalid operands}}
+
+  sl /= sl2;
+  ul /= ul2;
+  bl /= bl2; // expected-error {{invalid operands}}
+
+  fd /= fd2;
+
+  // -------------------------------------------------------------------------
+  // Test that / rules apply to % too, except that doubles are not allowed.
+  // -------------------------------------------------------------------------
+
+  sc = sc % sc2;
+  uc = uc % uc2;
+  bc = bc % bc2; // expected-error {{invalid operands}}
+
+  sc = uc % sc2; // expected-error {{can't convert}}
+  sc = sc % bc2; // expected-error {{can't convert}}
+  uc = bc % uc2; // expected-error {{can't convert}}
+
+  sc %= sc2;
+  uc %= uc2;
+  bc %= bc2; // expected-error {{invalid operands}}
+
+  sc %= uc2; // expected-error {{can't convert}}
+  uc %= bc2; // expected-error {{can't convert}}
+  bc %= sc2; // expected-error {{can't convert}}
+
+  ss %= ss2;
+  us %= us2;
+  bs %= bs2; // expected-error {{invalid operands}}
+
+  si %= si2;
+  ui %= ui2;
+  bi %= bi2; // expected-error {{invalid operands}}
+
+  sl %= sl2;
+  ul %= ul2;
+  bl %= bl2; // expected-error {{invalid operands}}
+
+  fd %= fd2; // expected-error {{invalid operands}}
+
+  // -------------------------------------------------------------------------
+  // Test &.
+  // -------------------------------------------------------------------------
+
+  sc = sc & sc2;
+  sc = sc & uc2; // expected-error {{can't convert}}
+  sc = uc & sc2; // expected-error {{can't convert}}
+  sc = sc & bc2;
+  sc = bc & sc2;
+
+  uc = uc & uc2;
+  uc = sc & uc2; // expected-error {{can't convert}}
+  uc = uc & sc2; // expected-error {{can't convert}}
+  uc = bc & uc2;
+  uc = uc & bc2;
+
+  bc = bc & bc2;
+  bc = bc & uc2; // expected-error {{incompatible type}}
+  bc = uc & bc2; // expected-error {{incompatible type}}
+  bc = bc & sc2; // expected-error {{incompatible type}}
+  bc = sc & bc2; // expected-error {{incompatible type}}
+
+  fd = fd & fd2; // expected-error {{invalid operands}}
+  fd = bl & fd2; // expected-error {{invalid operands}}
+  fd = fd & bl2; // expected-error {{invalid operands}}
+  fd = fd & sl2; // expected-error {{invalid operands}}
+  fd = fd & ul2; // expected-error {{invalid operands}}
+
+  sc &= sc2;
+  sc &= uc2; // expected-error {{can't convert}}
+  sc &= bc2;
+
+  uc &= uc2;
+  uc &= sc2; // expected-error {{can't convert}}
+  uc &= bc2;
+
+  bc &= bc2;
+  bc &= sc2; // expected-error {{can't convert}}
+  bc &= uc2; // expected-error {{can't convert}}
+
+  sc &= ss2; // expected-error {{can't convert}}
+  sc &= si2; // expected-error {{can't convert}}
+  sc &= sl2; // expected-error {{can't convert}}
+  sc &= fd2; // expected-error {{invalid operands}}
+
+  us &= bc2; // expected-error {{can't convert}}
+  ui &= bc2; // expected-error {{can't convert}}
+  ul &= bc2; // expected-error {{can't convert}}
+  fd &= bc2; // expected-error {{invalid operands}}
+
+  ss &= ss2;
+  us &= us2;
+  bs &= bs2;
+
+  si &= si2;
+  ui &= ui2;
+  bi &= bi2;
+
+  sl &= sl2;
+  ul &= ul2;
+  bl &= bl2;
+
+  // -------------------------------------------------------------------------
+  // Test that & rules apply to | too.
+  // -------------------------------------------------------------------------
+
+  sc = sc | sc2;
+  sc = sc | uc2; // expected-error {{can't convert}}
+  sc = sc | bc2;
+
+  uc = uc | uc2;
+  uc = sc | uc2; // expected-error {{can't convert}}
+  uc = bc | uc2;
+
+  bc = bc | bc2;
+  bc = uc | bc2; // expected-error {{incompatible type}}
+  bc = bc | sc2; // expected-error {{incompatible type}}
+
+  fd = fd | fd2; // expected-error {{invalid operands}}
+  fd = bl | fd2; // expected-error {{invalid operands}}
+
+  ss |= ss2;
+  us |= us2;
+  bs |= bs2;
+
+  si |= si2;
+  ui |= ui2;
+  bi |= bi2;
+
+  sl |= sl2;
+  ul |= ul2;
+  bl |= bl2;
+
+  fd |= bl2; // expected-error {{invalid operands}}
+  fd |= fd2; // expected-error {{invalid operands}}
+
+  // -------------------------------------------------------------------------
+  // Test that & rules apply to ^ too.
+  // -------------------------------------------------------------------------
+
+  sc = sc ^ sc2;
+  sc = sc ^ uc2; // expected-error {{can't convert}}
+  sc = sc ^ bc2;
+
+  uc = uc ^ uc2;
+  uc = sc ^ uc2; // expected-error {{can't convert}}
+  uc = bc ^ uc2;
+
+  bc = bc ^ bc2;
+  bc = uc ^ bc2; // expected-error {{incompatible type}}
+  bc = bc ^ sc2; // expected-error {{incompatible type}}
+
+  fd = fd ^ fd2; // expected-error {{invalid operands}}
+  fd = bl ^ fd2; // expected-error {{invalid operands}}
+
+  ss ^= ss2;
+  us ^= us2;
+  bs ^= bs2;
+
+  si ^= si2;
+  ui ^= ui2;
+  bi ^= bi2;
+
+  sl ^= sl2;
+  ul ^= ul2;
+  bl ^= bl2;
+
+  fd ^= bl2; // expected-error {{invalid operands}}
+  fd ^= fd2; // expected-error {{invalid operands}}
+
+  // -------------------------------------------------------------------------
+  // Test <<.
+  // -------------------------------------------------------------------------
+
+  sc = sc << sc2;
+  sc = sc << uc2;
+  sc = uc << sc2; // expected-error {{incompatible type}}
+  sc = sc << bc2; // expected-error {{invalid operands}}
+  sc = bc << sc2; // expected-error {{invalid operands}}
+
+  uc = uc << uc2;
+  uc = sc << uc2; // expected-error {{assigning to}}
+  uc = uc << sc2;
+  uc = bc << uc2; // expected-error {{invalid operands}}
+  uc = uc << bc2; // expected-error {{invalid operands}}
+
+  bc = bc << bc2; // expected-error {{invalid operands}}
+  bc = bc << uc2; // expected-error {{invalid operands}}
+  bc = uc << bc2; // expected-error {{invalid operands}}
+  bc = bc << sc2; // expected-error {{invalid operands}}
+  bc = sc << bc2; // expected-error {{invalid operands}}
+
+  sc = sc << 1;
+  sc = sc << 1.0f; // expected-error {{integer is required}}
+  sc = sc << sc_scalar;
+  sc = sc << uc_scalar;
+  sc = sc << ss_scalar;
+  sc = sc << us_scalar;
+  sc = sc << si_scalar;
+  sc = sc << ui_scalar;
+  sc = sc << sl_scalar;
+  sc = sc << ul_scalar;
+  sc = sc_scalar << sc; // expected-error {{first operand is not a vector}}
+  sc = uc_scalar << sc; // expected-error {{first operand is not a vector}}
+  uc = uc << sc_scalar;
+  uc = uc << uc_scalar;
+  uc = sc_scalar << uc; // expected-error {{first operand is not a vector}}
+  uc = uc_scalar << uc; // expected-error {{first operand is not a vector}}
+
+  ss = ss << ss2;
+  ss = ss << ss_scalar;
+  us = us << us2;
+  us = us << us_scalar;
+  bs = bs << bs2; // expected-error {{invalid operands}}
+
+  si = si << si2;
+  si = si << si_scalar;
+  ui = ui << ui2;
+  ui = ui << ui_scalar;
+  bi = bi << bi2; // expected-error {{invalid operands}}
+
+  sl = sl << sl2;
+  sl = sl << sl_scalar;
+  ul = ul << ul2;
+  ul = ul << ul_scalar;
+  bl = bl << bl2; // expected-error {{invalid operands}}
+
+  fd = fd << fd2; // expected-error {{integer is required}}
+  fd = fd << ul2; // expected-error {{integer is required}}
+  fd = sl << fd2; // expected-error {{integer is required}}
+
+  sc <<= sc2;
+  sc <<= uc2;
+  sc <<= bc2; // expected-error {{invalid operands}}
+  sc <<= sc_scalar;
+
+  uc <<= uc2;
+  uc <<= sc2;
+  uc <<= bc2; // expected-error {{invalid operands}}
+  uc <<= uc_scalar;
+
+  bc <<= bc2; // expected-error {{invalid operands}}
+  bc <<= sc2; // expected-error {{invalid operands}}
+  bc <<= uc2; // expected-error {{invalid operands}}
+
+  sc <<= ss2; // expected-error {{vector operands do not have the same number of elements}}
+  sc <<= si2; // expected-error {{vector operands do not have the same number of elements}}
+  sc <<= sl2; // expected-error {{vector operands do not have the same number of elements}}
+  sc <<= fd2; // expected-error {{integer is required}}
+
+  ss <<= ss2;
+  ss <<= ss_scalar;
+  us <<= us2;
+  us <<= us_scalar;
+  bs <<= bs2; // expected-error {{invalid operands}}
+
+  si <<= si2;
+  si <<= si_scalar;
+  ui <<= ui2;
+  ui <<= ui_scalar;
+  bi <<= bi2; // expected-error {{invalid operands}}
+
+  sl <<= sl2;
+  sl <<= sl_scalar;
+  ul <<= ul2;
+  ul <<= ul_scalar;
+  bl <<= bl2; // expected-error {{invalid operands}}
+
+  fd <<= fd2; // expected-error {{integer is required}}
+
+  // -------------------------------------------------------------------------
+  // Test >>.
+  // -------------------------------------------------------------------------
+
+  sc = sc >> sc2;
+  sc = sc >> uc2;
+  sc = uc >> sc2; // expected-error {{incompatible type}}
+  sc = sc >> bc2; // expected-error {{invalid operands}}
+  sc = bc >> sc2; // expected-error {{invalid operands}}
+
+  uc = uc >> uc2;
+  uc = sc >> uc2; // expected-error {{assigning to}}
+  uc = uc >> sc2;
+  uc = bc >> uc2; // expected-error {{invalid operands}}
+  uc = uc >> bc2; // expected-error {{invalid operands}}
+
+  bc = bc >> bc2; // expected-error {{invalid operands}}
+  bc = bc >> uc2; // expected-error {{invalid operands}}
+  bc = uc >> bc2; // expected-error {{invalid operands}}
+  bc = bc >> sc2; // expected-error {{invalid operands}}
+  bc = sc >> bc2; // expected-error {{invalid operands}}
+
+  sc = sc >> 1;
+  sc = sc >> 1.0f; // expected-error {{integer is required}}
+  sc = sc >> sc_scalar;
+  sc = sc >> uc_scalar;
+  sc = sc >> ss_scalar;
+  sc = sc >> us_scalar;
+  sc = sc >> si_scalar;
+  sc = sc >> ui_scalar;
+  sc = sc >> sl_scalar;
+  sc = sc >> ul_scalar;
+  sc = sc_scalar >> sc; // expected-error {{first operand is not a vector}}
+  sc = uc_scalar >> sc; // expected-error {{first operand is not a vector}}
+  uc = uc >> sc_scalar;
+  uc = uc >> uc_scalar;
+  uc = sc_scalar >> uc; // expected-error {{first operand is not a vector}}
+  uc = uc_scalar >> uc; // expected-error {{first operand is not a vector}}
+
+  ss = ss >> ss2;
+  ss = ss >> ss_scalar;
+  us = us >> us2;
+  us = us >> us_scalar;
+  bs = bs >> bs2; // expected-error {{invalid operands}}
+
+  si = si >> si2;
+  si = si >> si_scalar;
+  ui = ui >> ui2;
+  ui = ui >> ui_scalar;
+  bi = bi >> bi2; // expected-error {{invalid operands}}
+
+  sl = sl >> sl2;
+  sl = sl >> sl_scalar;
+  ul = ul >> ul2;
+  ul = ul >> ul_scalar;
+  bl = bl >> bl2; // expected-error {{invalid operands}}
+
+  fd = fd >> fd2; // expected-error {{integer is required}}
+  fd = fd >> ul2; // expected-error {{integer is required}}
+  fd = sl >> fd2; // expected-error {{integer is required}}
+
+  sc >>= sc2;
+  sc >>= uc2;
+  sc >>= bc2; // expected-error {{invalid operands}}
+  sc >>= sc_scalar;
+
+  uc >>= uc2;
+  uc >>= sc2;
+  uc >>= bc2; // expected-error {{invalid operands}}
+  uc >>= uc_scalar;
+
+  bc >>= bc2; // expected-error {{invalid operands}}
+  bc >>= sc2; // expected-error {{invalid operands}}
+  bc >>= uc2; // expected-error {{invalid operands}}
+
+  sc >>= ss2; // expected-error {{vector operands do not have the same number of elements}}
+  sc >>= si2; // expected-error {{vector operands do not have the same number of elements}}
+  sc >>= sl2; // expected-error {{vector operands do not have the same number of elements}}
+  sc >>= fd2; // expected-error {{integer is required}}
+
+  ss >>= ss2;
+  ss >>= ss_scalar;
+  us >>= us2;
+  us >>= us_scalar;
+  bs >>= bs2; // expected-error {{invalid operands}}
+
+  si >>= si2;
+  si >>= si_scalar;
+  ui >>= ui2;
+  ui >>= ui_scalar;
+  bi >>= bi2; // expected-error {{invalid operands}}
+
+  sl >>= sl2;
+  sl >>= sl_scalar;
+  ul >>= ul2;
+  ul >>= ul_scalar;
+  bl >>= bl2; // expected-error {{invalid operands}}
+
+  fd >>= fd2; // expected-error {{integer is required}}
+
+  // -------------------------------------------------------------------------
+  // Test ==.
+  // -------------------------------------------------------------------------
+
+  (void)(sc == sc2);
+  (void)(uc == uc2);
+  (void)(bc == bc2);
+
+  (void)(sc == uc); // expected-error {{can't convert}}
+  (void)(sc == bc);
+
+  (void)(uc == sc); // expected-error {{can't convert}}
+  (void)(uc == bc);
+
+  (void)(bc == sc);
+  (void)(bc == uc);
+
+  (void)(ss == ss2);
+  (void)(us == us2);
+  (void)(bs == bs2);
+
+  (void)(si == si2);
+  (void)(ui == ui2);
+  (void)(bi == bi2);
+
+  (void)(sl == sl2);
+  (void)(ul == ul2);
+  (void)(bl == bl2);
+  (void)(fd == fd2);
+
+  (void)(fd == ul); // expected-error {{can't convert}}
+  (void)(ul == fd); // expected-error {{can't convert}}
+
+  // -------------------------------------------------------------------------
+  // Test that == rules apply to != too.
+  // -------------------------------------------------------------------------
+
+  (void)(sc != sc2);
+  (void)(uc != uc2);
+  (void)(bc != bc2);
+
+  (void)(sc != uc); // expected-error {{can't convert}}
+  (void)(sc != bc);
+
+  (void)(ss != ss2);
+  (void)(us != us2);
+  (void)(bs != bs2);
+
+  (void)(si != si2);
+  (void)(ui != ui2);
+  (void)(bi != bi2);
+
+  (void)(sl != sl2);
+  (void)(ul != ul2);
+  (void)(bl != bl2);
+  (void)(fd != fd2);
+
+  // -------------------------------------------------------------------------
+  // Test that == rules apply to <= too.
+  // -------------------------------------------------------------------------
+
+  (void)(sc <= sc2);
+  (void)(uc <= uc2);
+  (void)(bc <= bc2);
+
+  (void)(sc <= uc); // expected-error {{can't convert}}
+  (void)(sc <= bc);
+
+  (void)(ss <= ss2);
+  (void)(us <= us2);
+  (void)(bs <= bs2);
+
+  (void)(si <= si2);
+  (void)(ui <= ui2);
+  (void)(bi <= bi2);
+
+  (void)(sl <= sl2);
+  (void)(ul <= ul2);
+  (void)(bl <= bl2);
+  (void)(fd <= fd2);
+
+  // -------------------------------------------------------------------------
+  // Test that == rules apply to >= too.
+  // -------------------------------------------------------------------------
+
+  (void)(sc >= sc2);
+  (void)(uc >= uc2);
+  (void)(bc >= bc2);
+
+  (void)(sc >= uc); // expected-error {{can't convert}}
+  (void)(sc >= bc);
+
+  (void)(ss >= ss2);
+  (void)(us >= us2);
+  (void)(bs >= bs2);
+
+  (void)(si >= si2);
+  (void)(ui >= ui2);
+  (void)(bi >= bi2);
+
+  (void)(sl >= sl2);
+  (void)(ul >= ul2);
+  (void)(bl >= bl2);
+  (void)(fd >= fd2);
+
+  // -------------------------------------------------------------------------
+  // Test that == rules apply to < too.
+  // -------------------------------------------------------------------------
+
+  (void)(sc < sc2);
+  (void)(uc < uc2);
+  (void)(bc < bc2);
+
+  (void)(sc < uc); // expected-error {{can't convert}}
+  (void)(sc < bc);
+
+  (void)(ss < ss2);
+  (void)(us < us2);
+  (void)(bs < bs2);
+
+  (void)(si < si2);
+  (void)(ui < ui2);
+  (void)(bi < bi2);
+
+  (void)(sl < sl2);
+  (void)(ul < ul2);
+  (void)(bl < bl2);
+  (void)(fd < fd2);
+
+  // -------------------------------------------------------------------------
+  // Test that == rules apply to > too.
+  // -------------------------------------------------------------------------
+
+  (void)(sc > sc2);
+  (void)(uc > uc2);
+  (void)(bc > bc2);
+
+  (void)(sc > uc); // expected-error {{can't convert}}
+  (void)(sc > bc);
+
+  (void)(ss > ss2);
+  (void)(us > us2);
+  (void)(bs > bs2);
+
+  (void)(si > si2);
+  (void)(ui > ui2);
+  (void)(bi > bi2);
+
+  (void)(sl > sl2);
+  (void)(ul > ul2);
+  (void)(bl > bl2);
+  (void)(fd > fd2);
+}