[PowerPC] Provide some P8-specific altivec overloads for P7
This adds additional support for XL compatibility. There are a number
of functions in altivec.h that produce a single instruction (or a
very short sequence) for Power8 but can be done on Power7 without
scalarization. XL provides these implementations.
This patch adds the following overloads for doubleword vectors:
vec_add
vec_cmpeq
vec_cmpgt
vec_cmpge
vec_cmplt
vec_cmple
vec_sl
vec_sr
vec_sra
diff --git a/clang/lib/Headers/altivec.h b/clang/lib/Headers/altivec.h
index cb4f35c..0441c57 100644
--- a/clang/lib/Headers/altivec.h
+++ b/clang/lib/Headers/altivec.h
@@ -309,6 +309,26 @@
vec_add_u128(vector unsigned char __a, vector unsigned char __b) {
return __builtin_altivec_vadduqm(__a, __b);
}
+#elif defined(__VSX__)
+static __inline__ vector signed long long __ATTRS_o_ai
+vec_add(vector signed long long __a, vector signed long long __b) {
+ vector unsigned int __res =
+ (vector unsigned int)__a + (vector unsigned int)__b;
+ vector unsigned int __carry = __builtin_altivec_vaddcuw(
+ (vector unsigned int)__a, (vector unsigned int)__b);
+#ifdef __LITTLE_ENDIAN__
+ __carry = __builtin_shufflevector(__carry, __carry, 3, 0, 1, 2);
+#else
+ __carry = __builtin_shufflevector(__carry, __carry, 1, 2, 3, 0);
+#endif
+ return (vector signed long long)(__res + __carry);
+}
+
+static __inline__ vector unsigned long long __ATTRS_o_ai
+vec_add(vector unsigned long long __a, vector unsigned long long __b) {
+ return (vector unsigned long long)vec_add((vector signed long long)__a,
+ (vector signed long long)__b);
+}
#endif // defined(__POWER8_VECTOR__) && defined(__powerpc64__)
static __inline__ vector float __ATTRS_o_ai vec_add(vector float __a,
@@ -1730,7 +1750,31 @@
return (vector bool long long)__builtin_altivec_vcmpequd(
(vector long long)__a, (vector long long)__b);
}
+#else
+static __inline__ vector bool long long __ATTRS_o_ai
+vec_cmpeq(vector signed long long __a, vector signed long long __b) {
+ vector bool int __wordcmp =
+ vec_cmpeq((vector signed int)__a, (vector signed int)__b);
+#ifdef __LITTLE_ENDIAN__
+ __wordcmp &= __builtin_shufflevector(__wordcmp, __wordcmp, 3, 0, 1, 2);
+ return (vector bool long long)__builtin_shufflevector(__wordcmp, __wordcmp, 1,
+ 1, 3, 3);
+#else
+ __wordcmp &= __builtin_shufflevector(__wordcmp, __wordcmp, 1, 2, 3, 0);
+ return (vector bool long long)__builtin_shufflevector(__wordcmp, __wordcmp, 0,
+ 0, 2, 2);
+#endif
+}
+static __inline__ vector bool long long __ATTRS_o_ai
+vec_cmpeq(vector unsigned long long __a, vector unsigned long long __b) {
+ return vec_cmpeq((vector signed long long)__a, (vector signed long long)__b);
+}
+
+static __inline__ vector bool long long __ATTRS_o_ai
+vec_cmpeq(vector bool long long __a, vector bool long long __b) {
+ return vec_cmpeq((vector signed long long)__a, (vector signed long long)__b);
+}
#endif
static __inline__ vector bool int __ATTRS_o_ai vec_cmpeq(vector float __a,
@@ -2018,6 +2062,24 @@
return (vector bool long long)
~(__builtin_altivec_vcmpequd((vector long long)__a, (vector long long)__b));
}
+#else
+static __inline__ vector bool long long __ATTRS_o_ai
+vec_cmpne(vector bool long long __a, vector bool long long __b) {
+ return (vector bool long long)~(
+ vec_cmpeq((vector signed long long)__a, (vector signed long long)__b));
+}
+
+static __inline__ vector bool long long __ATTRS_o_ai
+vec_cmpne(vector signed long long __a, vector signed long long __b) {
+ return (vector bool long long)~(
+ vec_cmpeq((vector signed long long)__a, (vector signed long long)__b));
+}
+
+static __inline__ vector bool long long __ATTRS_o_ai
+vec_cmpne(vector unsigned long long __a, vector unsigned long long __b) {
+ return (vector bool long long)~(
+ vec_cmpeq((vector signed long long)__a, (vector signed long long)__b));
+}
#endif
#ifdef __VSX__
@@ -2070,6 +2132,46 @@
vec_cmpgt(vector unsigned long long __a, vector unsigned long long __b) {
return (vector bool long long)__builtin_altivec_vcmpgtud(__a, __b);
}
+#else
+static __inline__ vector bool long long __ATTRS_o_ai
+vec_cmpgt(vector signed long long __a, vector signed long long __b) {
+ vector signed int __sgtw =
+ vec_cmpgt((vector signed int)__a, (vector signed int)__b);
+ vector unsigned int __ugtw =
+ vec_cmpgt((vector unsigned int)__a, (vector unsigned int)__b);
+ vector bool int __eqw =
+ vec_cmpeq((vector signed int)__a, (vector signed int)__b);
+#ifdef __LITTLE_ENDIAN__
+ __ugtw = __builtin_shufflevector(__ugtw, __ugtw, 3, 0, 1, 2) & __eqw;
+ __sgtw |= (vector signed int)__ugtw;
+ return (vector bool long long)__builtin_shufflevector(__sgtw, __sgtw, 1, 1, 3,
+ 3);
+#else
+ __ugtw = __builtin_shufflevector(__ugtw, __ugtw, 1, 2, 3, 0) & __eqw;
+ __sgtw |= (vector signed int)__ugtw;
+ return (vector bool long long)__builtin_shufflevector(__sgtw, __sgtw, 0, 0, 2,
+ 2);
+#endif
+}
+
+static __inline__ vector bool long long __ATTRS_o_ai
+vec_cmpgt(vector unsigned long long __a, vector unsigned long long __b) {
+ vector unsigned int __ugtw =
+ vec_cmpgt((vector unsigned int)__a, (vector unsigned int)__b);
+ vector bool int __eqw =
+ vec_cmpeq((vector signed int)__a, (vector signed int)__b);
+#ifdef __LITTLE_ENDIAN__
+ __eqw = __builtin_shufflevector(__ugtw, __ugtw, 3, 0, 1, 2) & __eqw;
+ __ugtw |= __eqw;
+ return (vector bool long long)__builtin_shufflevector(__ugtw, __ugtw, 1, 1, 3,
+ 3);
+#else
+ __eqw = __builtin_shufflevector(__ugtw, __ugtw, 1, 2, 3, 0) & __eqw;
+ __ugtw |= __eqw;
+ return (vector bool long long)__builtin_shufflevector(__ugtw, __ugtw, 0, 0, 2,
+ 2);
+#endif
+}
#endif
static __inline__ vector bool int __ATTRS_o_ai vec_cmpgt(vector float __a,
@@ -2150,7 +2252,6 @@
}
#endif
-#ifdef __POWER8_VECTOR__
static __inline__ vector bool long long __ATTRS_o_ai
vec_cmpge(vector signed long long __a, vector signed long long __b) {
return ~(vec_cmpgt(__b, __a));
@@ -2160,7 +2261,6 @@
vec_cmpge(vector unsigned long long __a, vector unsigned long long __b) {
return ~(vec_cmpgt(__b, __a));
}
-#endif
#if defined(__POWER10_VECTOR__) && defined(__SIZEOF_INT128__)
static __inline__ vector bool __int128 __ATTRS_o_ai
@@ -2274,7 +2374,6 @@
}
#endif
-#ifdef __POWER8_VECTOR__
static __inline__ vector bool long long __ATTRS_o_ai
vec_cmple(vector signed long long __a, vector signed long long __b) {
return vec_cmpge(__b, __a);
@@ -2284,7 +2383,6 @@
vec_cmple(vector unsigned long long __a, vector unsigned long long __b) {
return vec_cmpge(__b, __a);
}
-#endif
#if defined(__POWER10_VECTOR__) && defined(__SIZEOF_INT128__)
static __inline__ vector bool __int128 __ATTRS_o_ai
@@ -2354,7 +2452,6 @@
}
#endif
-#ifdef __POWER8_VECTOR__
static __inline__ vector bool long long __ATTRS_o_ai
vec_cmplt(vector signed long long __a, vector signed long long __b) {
return vec_cmpgt(__b, __a);
@@ -2365,6 +2462,7 @@
return vec_cmpgt(__b, __a);
}
+#ifdef __POWER8_VECTOR__
/* vec_popcnt */
static __inline__ vector signed char __ATTRS_o_ai
@@ -8729,6 +8827,46 @@
vec_sl(vector long long __a, vector unsigned long long __b) {
return (vector long long)vec_sl((vector unsigned long long)__a, __b);
}
+#else
+static __inline__ vector unsigned char __ATTRS_o_ai
+vec_vspltb(vector unsigned char __a, unsigned char __b);
+static __inline__ vector unsigned long long __ATTRS_o_ai
+vec_sl(vector unsigned long long __a, vector unsigned long long __b) {
+ __b %= (vector unsigned long long)(sizeof(unsigned long long) * __CHAR_BIT__);
+
+ // Big endian element one (the right doubleword) can be left shifted as-is.
+ // The other element needs to be swapped into the right doubleword and
+ // shifted. Then the right doublewords of the two result vectors are merged.
+ vector signed long long __rightelt =
+ (vector signed long long)__builtin_altivec_vslo((vector signed int)__a,
+ (vector signed int)__b);
+#ifdef __LITTLE_ENDIAN__
+ __rightelt = (vector signed long long)__builtin_altivec_vsl(
+ (vector signed int)__rightelt, vec_vspltb((vector unsigned char)__b, 0));
+#else
+ __rightelt = (vector signed long long)__builtin_altivec_vsl(
+ (vector signed int)__rightelt, vec_vspltb((vector unsigned char)__b, 15));
+#endif
+ __a = __builtin_shufflevector(__a, __a, 1, 0);
+ __b = __builtin_shufflevector(__b, __b, 1, 0);
+ vector signed long long __leftelt =
+ (vector signed long long)__builtin_altivec_vslo((vector signed int)__a,
+ (vector signed int)__b);
+#ifdef __LITTLE_ENDIAN__
+ __leftelt = (vector signed long long)__builtin_altivec_vsl(
+ (vector signed int)__leftelt, vec_vspltb((vector unsigned char)__b, 0));
+ return __builtin_shufflevector(__rightelt, __leftelt, 0, 2);
+#else
+ __leftelt = (vector signed long long)__builtin_altivec_vsl(
+ (vector signed int)__leftelt, vec_vspltb((vector unsigned char)__b, 15));
+ return __builtin_shufflevector(__leftelt, __rightelt, 1, 3);
+#endif
+}
+
+static __inline__ vector long long __ATTRS_o_ai
+vec_sl(vector long long __a, vector unsigned long long __b) {
+ return (vector long long)vec_sl((vector unsigned long long)__a, __b);
+}
#endif
/* vec_vslb */
@@ -10194,6 +10332,47 @@
vec_sr(vector long long __a, vector unsigned long long __b) {
return (vector long long)vec_sr((vector unsigned long long)__a, __b);
}
+#else
+static __inline__ vector unsigned long long __ATTRS_o_ai
+vec_sr(vector unsigned long long __a, vector unsigned long long __b) {
+ __b %= (vector unsigned long long)(sizeof(unsigned long long) * __CHAR_BIT__);
+
+ // Big endian element zero (the left doubleword) can be right shifted as-is.
+ // However the shift amount must be in the right doubleword.
+ // The other element needs to be swapped into the left doubleword and
+ // shifted. Then the left doublewords of the two result vectors are merged.
+ vector signed long long __swapshift = __builtin_shufflevector(__b, __b, 1, 0);
+ vector signed long long __leftelt =
+ (vector signed long long)__builtin_altivec_vsro((vector signed int)__a,
+ __swapshift);
+#ifdef __LITTLE_ENDIAN__
+ __leftelt = (vector signed long long)__builtin_altivec_vsr(
+ (vector signed int)__leftelt,
+ vec_vspltb((vector unsigned char)__swapshift, 0));
+#else
+ __leftelt = (vector signed long long)__builtin_altivec_vsr(
+ (vector signed int)__leftelt,
+ vec_vspltb((vector unsigned char)__swapshift, 15));
+#endif
+ __a = __builtin_shufflevector(__a, __a, 1, 0);
+ vector signed long long __rightelt =
+ (vector signed long long)__builtin_altivec_vsro((vector signed int)__a,
+ (vector signed int)__b);
+#ifdef __LITTLE_ENDIAN__
+ __rightelt = (vector signed long long)__builtin_altivec_vsr(
+ (vector signed int)__rightelt, vec_vspltb((vector unsigned char)__b, 0));
+ return __builtin_shufflevector(__rightelt, __leftelt, 1, 3);
+#else
+ __rightelt = (vector signed long long)__builtin_altivec_vsr(
+ (vector signed int)__rightelt, vec_vspltb((vector unsigned char)__b, 15));
+ return __builtin_shufflevector(__leftelt, __rightelt, 0, 2);
+#endif
+}
+
+static __inline__ vector long long __ATTRS_o_ai
+vec_sr(vector long long __a, vector unsigned long long __b) {
+ return (vector long long)vec_sr((vector unsigned long long)__a, __b);
+}
#endif
/* vec_vsrb */
@@ -10280,6 +10459,18 @@
vec_sra(vector unsigned long long __a, vector unsigned long long __b) {
return (vector unsigned long long)((vector signed long long)__a >> __b);
}
+#else
+static __inline__ vector signed long long __ATTRS_o_ai
+vec_sra(vector signed long long __a, vector unsigned long long __b) {
+ __b %= (vector unsigned long long)(sizeof(unsigned long long) * __CHAR_BIT__);
+ return __a >> __b;
+}
+
+static __inline__ vector unsigned long long __ATTRS_o_ai
+vec_sra(vector unsigned long long __a, vector unsigned long long __b) {
+ __b %= (vector unsigned long long)(sizeof(unsigned long long) * __CHAR_BIT__);
+ return (vector unsigned long long)((vector signed long long)__a >> __b);
+}
#endif
/* vec_vsrab */
diff --git a/clang/test/CodeGen/builtins-ppc-vsx.c b/clang/test/CodeGen/builtins-ppc-vsx.c
index ecae9a6..4bb7a05 100644
--- a/clang/test/CodeGen/builtins-ppc-vsx.c
+++ b/clang/test/CodeGen/builtins-ppc-vsx.c
@@ -2313,3 +2313,282 @@
// CHECK-LE: fdiv fast <2 x double> <double 1.000000e+00, double 1.000000e+00>
return vec_rsqrt(a);
}
+
+void test_p8overloads_backwards_compat() {
+ // CHECK-LABEL: test_p8overloads_backwards_compat
+ res_vsll = vec_add(vsll, vsll);
+ // CHECK: add <4 x i32>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vaddcuw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: add <4 x i32>
+ // CHECK-LE: add <4 x i32>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vaddcuw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: add <4 x i32>
+ res_vull = vec_add(vull, vull);
+ // CHECK: add <4 x i32>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vaddcuw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: add <4 x i32>
+ // CHECK-LE: add <4 x i32>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vaddcuw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: add <4 x i32>
+ dummy();
+ // CHECK: call void @dummy()
+ // CHECK-LE: call void @dummy()
+
+ res_vbll = vec_cmpeq(vsll, vsll);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ res_vbll = vec_cmpeq(vull, vull);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ res_vbll = vec_cmpeq(vbll, vbll);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ dummy();
+ // CHECK: call void @dummy()
+ // CHECK-LE: call void @dummy()
+
+ res_vbll = vec_cmpgt(vsll, vsll);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtsw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: or <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtsw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: or <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ res_vbll = vec_cmpgt(vull, vull);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: or <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: or <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ dummy();
+ // CHECK: call void @dummy()
+ // CHECK-LE: call void @dummy()
+
+ res_vbll = vec_cmpge(vsll, vsll);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtsw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: or <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK: xor <2 x i64> {{%.*}}, <i64 -1, i64 -1>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtsw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: or <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ // CHECK-LE: xor <2 x i64> {{%.*}}, <i64 -1, i64 -1>
+ res_vbll = vec_cmpge(vull, vull);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: or <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK: xor <2 x i64> {{%.*}}, <i64 -1, i64 -1>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: or <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ // CHECK-LE: xor <2 x i64> {{%.*}}, <i64 -1, i64 -1>
+ dummy();
+ // CHECK: call void @dummy()
+ // CHECK-LE: call void @dummy()
+
+ res_vbll = vec_cmplt(vsll, vsll);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtsw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: or <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtsw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: or <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ res_vbll = vec_cmplt(vull, vull);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: or <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: or <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ dummy();
+ // CHECK: call void @dummy()
+ // CHECK-LE: call void @dummy()
+
+ res_vbll = vec_cmple(vsll, vsll);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtsw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: or <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK: xor <2 x i64> {{%.*}}, <i64 -1, i64 -1>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtsw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: or <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ // CHECK-LE: xor <2 x i64> {{%.*}}, <i64 -1, i64 -1>
+ res_vbll = vec_cmple(vull, vull);
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 2, i32 3, i32 0>
+ // CHECK: and <4 x i32>
+ // CHECK: or <4 x i32>
+ // CHECK: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
+ // CHECK: xor <2 x i64> {{%.*}}, <i64 -1, i64 -1>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpgtuw
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vcmpequw
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 3, i32 0, i32 1, i32 2>
+ // CHECK-LE: and <4 x i32>
+ // CHECK-LE: or <4 x i32>
+ // CHECK-LE: shufflevector <4 x i32> {{%.*}}, <4 x i32> {{%.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
+ // CHECK-LE: xor <2 x i64> {{%.*}}, <i64 -1, i64 -1>
+ dummy();
+ // CHECK: call void @dummy()
+ // CHECK-LE: call void @dummy()
+
+ res_vsll = vec_sl(vsll, vull);
+ // CHECK: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vslo
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsl
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vslo
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsl
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 3>
+ // CHECK-LE: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vslo
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsl
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vslo
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsl
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 0, i32 2>
+ res_vull = vec_sl(vull, vull);
+ // CHECK: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vslo
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsl
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vslo
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsl
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 3>
+ // CHECK-LE: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vslo
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsl
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vslo
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsl
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 0, i32 2>
+ dummy();
+ // CHECK: call void @dummy()
+ // CHECK-LE: call void @dummy()
+
+ res_vsll = vec_sr(vsll, vull);
+ // CHECK: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsro
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsr
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsro
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsr
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 0, i32 2>
+ // CHECK-LE: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsro
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsr
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsro
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsr
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 3>
+ res_vull = vec_sr(vull, vull);
+ // CHECK: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsro
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsr
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsro
+ // CHECK: call <4 x i32> @llvm.ppc.altivec.vsr
+ // CHECK: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 0, i32 2>
+ // CHECK-LE: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsro
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsr
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 0>
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsro
+ // CHECK-LE: call <4 x i32> @llvm.ppc.altivec.vsr
+ // CHECK-LE: shufflevector <2 x i64> {{%.*}}, <2 x i64> {{%.*}}, <2 x i32> <i32 1, i32 3>
+ dummy();
+ // CHECK: call void @dummy()
+ // CHECK-LE: call void @dummy()
+
+ res_vsll = vec_sra(vsll, vull);
+ // CHECK: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK: ashr <2 x i64>
+ // CHECK-LE: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK-LE: ashr <2 x i64>
+ res_vull = vec_sra(vull, vull);
+ // CHECK: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK: ashr <2 x i64>
+ // CHECK-LE: urem <2 x i64> {{%.*}}, <i64 64, i64 64>
+ // CHECK-LE: ashr <2 x i64>
+}
