llvm/test/Transforms/LoopVectorize/constantfolder.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 5
; RUN: opt -passes=loop-vectorize -force-vector-width=4 -S %s | FileCheck %s

define void @const_fold_ptradd(ptr %dst, i64 %d) {
; CHECK-LABEL: define void @const_fold_ptradd(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store i16 0, ptr [[DST]], align 2
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP1:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP1]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:                                         ; preds = %loop.latch, %entry
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 true, label %loop.latch, label %else

else:
  br label %loop.latch

loop.latch:
  %const.0 = phi i64 [ %d, %else ], [ 0, %loop.header ]
  %gep = getelementptr i16, ptr %dst, i64 %const.0
  store i16 0, ptr %gep, align 2
  %iv.next = add i64 %iv, 1
  %cmp = icmp ult i64 %iv.next, 100
  br i1 %cmp, label %loop.header, label %exit

exit:
  ret void
}

define void @const_fold_inbounds_ptradd(ptr %dst, i64 %d) {
; CHECK-LABEL: define void @const_fold_inbounds_ptradd(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store i16 0, ptr [[DST]], align 2
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP1:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP1]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:                                         ; preds = %loop.latch, %entry
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 true, label %loop.latch, label %else

else:
  br label %loop.latch

loop.latch:
  %const.0 = phi i64 [ %d, %else ], [ 0, %loop.header ]
  %gep = getelementptr inbounds i16, ptr %dst, i64 %const.0
  store i16 0, ptr %gep, align 2
  %iv.next = add i64 %iv, 1
  %cmp = icmp ult i64 %iv.next, 100
  br i1 %cmp, label %loop.header, label %exit

exit:
  ret void
}

define void @const_fold_select(ptr %dst, i64 %d) {
; CHECK-LABEL: define void @const_fold_select(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    [[TMP3:%.*]] = or i64 [[D]], 1
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store i64 [[TMP3]], ptr [[DST]], align 8
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP4]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:                                         ; preds = %loop.latch, %entry
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 true, label %loop.latch, label %else

else:
  br label %loop.latch

loop.latch:
  %const.1 = phi i64 [ %d, %else ], [ 1, %loop.header ]
  %or = or i64 %d, %const.1
  store i64 %or, ptr %dst, align 8
  %iv.next = add i64 %iv, 1
  %cmp = icmp ult i64 %iv.next, 100
  br i1 %cmp, label %loop.header, label %exit

exit:
  ret void
}

define void @const_fold_add_sub_mul_ashr_lshr(ptr %dst, i64 %d) {
; CHECK-LABEL: define void @const_fold_add_sub_mul_ashr_lshr(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store i64 1, ptr [[DST]], align 8
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP6]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:                                         ; preds = %loop.latch, %entry
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 true, label %loop.latch, label %else

else:
  br label %loop.latch

loop.latch:
  %const.1 = phi i64 [ %d, %else ], [ 1, %loop.header ]
  %add = add i64 2, %const.1
  %sub = sub i64 %add, %const.1
  %ashr = ashr i64 %sub, %const.1
  %mul = mul i64 %ashr, 3
  %lshr = lshr i64 %mul, %const.1
  store i64 %lshr, ptr %dst, align 8
  %iv.next = add i64 %iv, 1
  %cmp = icmp ult i64 %iv.next, 100
  br i1 %cmp, label %loop.header, label %exit

exit:
  ret void
}

define void @const_fold_and_or_xor(ptr %dst, i64 %d) {
; CHECK-LABEL: define void @const_fold_and_or_xor(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store i64 1, ptr [[DST]], align 8
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP4]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:                                         ; preds = %loop.latch, %entry
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 true, label %loop.latch, label %else

else:
  br label %loop.latch

loop.latch:
  %const.1 = phi i64 [ %d, %else ], [ 1, %loop.header ]
  %or = or i64 2, %const.1
  %and = and i64 %or, %const.1
  %xor = and i64 %and, %const.1
  store i64 %xor, ptr %dst, align 8
  %iv.next = add i64 %iv, 1
  %cmp = icmp ult i64 %iv.next, 100
  br i1 %cmp, label %loop.header, label %exit

exit:
  ret void
}

define void @const_fold_cmp_zext(ptr %dst, i64 %d) {
; CHECK-LABEL: define void @const_fold_cmp_zext(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store i8 1, ptr [[DST]], align 1
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP3:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP3]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:                                         ; preds = %loop.latch, %entry
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 true, label %loop.latch, label %else

else:
  br label %loop.latch

loop.latch:
  %const.1 = phi i64 [ %d, %else ], [ 1, %loop.header ]
  %val = icmp ugt i64 2, %const.1
  %zext = zext i1 %val to i8
  store i8 %zext, ptr %dst, align 1
  %iv.next = add i64 %iv, 1
  %cmp = icmp ult i64 %iv.next, 100
  br i1 %cmp, label %loop.header, label %exit

exit:
  ret void
}

define void @const_fold_trunc(ptr %dst, i64 %d) {
; CHECK-LABEL: define void @const_fold_trunc(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store i16 0, ptr [[DST]], align 2
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP2:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP2]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:                                         ; preds = %loop.latch, %entry
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 true, label %loop.latch, label %else

else:
  br label %loop.latch

loop.latch:
  %const.0 = phi i64 [ %d, %else ], [ 0, %loop.header ]
  %trunc = trunc i64 %const.0 to i16
  store i16 %trunc, ptr %dst, align 2
  %iv.next = add i64 %iv, 1
  %cmp = icmp ult i64 %iv.next, 100
  br i1 %cmp, label %loop.header, label %exit

exit:
  ret void
}

define void @const_fold_binaryintrinsic(ptr %dst, i64 %d) {
; CHECK-LABEL: define void @const_fold_binaryintrinsic(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store i64 3, ptr [[DST]], align 2
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP2:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP2]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP9:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %const.0 = xor i64 %d, %d
  %trunc = call i64 @llvm.umax.i64(i64 %const.0, i64 3)
  store i64 %trunc, ptr %dst, align 2
  %iv.next = add i64 %iv, 1
  %cmp = icmp ult i64 %iv.next, 100
  br i1 %cmp, label %loop, label %exit

exit:
  ret void
}

define void @const_fold_widegep(ptr noalias %A, ptr noalias %B, i64 %d) {
; CHECK-LABEL: define void @const_fold_widegep(
; CHECK-SAME: ptr noalias [[A:%.*]], ptr noalias [[B:%.*]], i64 [[D:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    store ptr [[A]], ptr [[B]], align 8
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP0:%.*]] = icmp eq i64 [[INDEX_NEXT]], 100
; CHECK-NEXT:    br i1 [[TMP0]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %const.0 = xor i64 %d, %d
  %gep.A = getelementptr i64, ptr %A, i64 %const.0
  %gep.B = getelementptr i64, ptr %B, i64 %const.0
  store ptr %gep.A, ptr %gep.B
  %iv.next = add nuw nsw i64 %iv, 1
  %exit.cond = icmp ult i64 %iv.next, 100
  br i1 %exit.cond, label %loop, label %exit

exit:
  ret void
}

; The test intentionally uses a large number of live-ins to guard against a
; use-after-free when reallocating
; (https://github.com/llvm/llvm-project/issues/173222).
define void @unused_live_ins_in_preheader(ptr %dst, ptr %src) {
; CHECK-LABEL: define void @unused_live_ins_in_preheader(
; CHECK-SAME: ptr [[DST:%.*]], ptr [[SRC:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    [[CMP1:%.*]] = icmp ne i8 0, 0
; CHECK-NEXT:    [[ZEXT1:%.*]] = zext i1 false to i32
; CHECK-NEXT:    [[TRUNC1:%.*]] = trunc i64 0 to i16
; CHECK-NEXT:    [[CMP2:%.*]] = icmp sgt i16 0, 0
; CHECK-NEXT:    [[SEL1:%.*]] = select i1 false, i16 0, i16 0
; CHECK-NEXT:    [[ZEXT2:%.*]] = zext i16 0 to i32
; CHECK-NEXT:    [[LSHR1:%.*]] = lshr i32 0, 0
; CHECK-NEXT:    [[TRUNC2:%.*]] = trunc i32 0 to i16
; CHECK-NEXT:    [[XOR1:%.*]] = xor i16 0, 0
; CHECK-NEXT:    [[ZEXT3:%.*]] = zext i16 0 to i32
; CHECK-NEXT:    [[LD:%.*]] = load i32, ptr [[SRC]], align 4
; CHECK-NEXT:    [[CMP3:%.*]] = icmp sgt i32 0, 0
; CHECK-NEXT:    [[LSHR2:%.*]] = lshr i32 0, 0
; CHECK-NEXT:    [[CMP4:%.*]] = icmp ne i32 0, 0
; CHECK-NEXT:    [[SEL2:%.*]] = select i1 false, i1 false, i1 false
; CHECK-NEXT:    [[SHL1:%.*]] = shl i32 0, 0
; CHECK-NEXT:    [[SEL3:%.*]] = select i1 false, i32 0, i32 0
; CHECK-NEXT:    [[SEXT1:%.*]] = sext i8 0 to i32
; CHECK-NEXT:    [[CMP5:%.*]] = icmp eq i32 0, 0
; CHECK-NEXT:    [[ZEXT4:%.*]] = zext i32 0 to i64
; CHECK-NEXT:    [[SUB1:%.*]] = sub i64 0, 0
; CHECK-NEXT:    [[SEXT2:%.*]] = sext i8 0 to i64
; CHECK-NEXT:    [[AND1:%.*]] = and i64 0, 0
; CHECK-NEXT:    [[CMP6:%.*]] = icmp ne i64 0, 0
; CHECK-NEXT:    [[ZEXT5:%.*]] = zext i1 false to i16
; CHECK-NEXT:    [[CMP7:%.*]] = icmp eq i16 0, 0
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i32 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[PRED_STORE_CONTINUE6:.*]] ]
; CHECK-NEXT:    [[VEC_IND:%.*]] = phi <4 x i8> [ <i8 0, i8 1, i8 2, i8 3>, %[[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], %[[PRED_STORE_CONTINUE6]] ]
; CHECK-NEXT:    [[TMP0:%.*]] = icmp ule <4 x i8> [[VEC_IND]], splat (i8 8)
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x i1> [[TMP0]], i32 0
; CHECK-NEXT:    br i1 [[TMP1]], label %[[PRED_STORE_IF:.*]], label %[[PRED_STORE_CONTINUE:.*]]
; CHECK:       [[PRED_STORE_IF]]:
; CHECK-NEXT:    [[TMP2:%.*]] = add i32 [[INDEX]], 0
; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr i32, ptr [[DST]], i32 [[TMP2]]
; CHECK-NEXT:    store i32 [[TMP2]], ptr [[TMP3]], align 8
; CHECK-NEXT:    br label %[[PRED_STORE_CONTINUE]]
; CHECK:       [[PRED_STORE_CONTINUE]]:
; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <4 x i1> [[TMP0]], i32 1
; CHECK-NEXT:    br i1 [[TMP4]], label %[[PRED_STORE_IF1:.*]], label %[[PRED_STORE_CONTINUE2:.*]]
; CHECK:       [[PRED_STORE_IF1]]:
; CHECK-NEXT:    [[TMP5:%.*]] = add i32 [[INDEX]], 1
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr i32, ptr [[DST]], i32 [[TMP5]]
; CHECK-NEXT:    store i32 [[TMP5]], ptr [[TMP6]], align 8
; CHECK-NEXT:    br label %[[PRED_STORE_CONTINUE2]]
; CHECK:       [[PRED_STORE_CONTINUE2]]:
; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <4 x i1> [[TMP0]], i32 2
; CHECK-NEXT:    br i1 [[TMP7]], label %[[PRED_STORE_IF3:.*]], label %[[PRED_STORE_CONTINUE4:.*]]
; CHECK:       [[PRED_STORE_IF3]]:
; CHECK-NEXT:    [[TMP8:%.*]] = add i32 [[INDEX]], 2
; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr i32, ptr [[DST]], i32 [[TMP8]]
; CHECK-NEXT:    store i32 [[TMP8]], ptr [[TMP9]], align 8
; CHECK-NEXT:    br label %[[PRED_STORE_CONTINUE4]]
; CHECK:       [[PRED_STORE_CONTINUE4]]:
; CHECK-NEXT:    [[TMP10:%.*]] = extractelement <4 x i1> [[TMP0]], i32 3
; CHECK-NEXT:    br i1 [[TMP10]], label %[[PRED_STORE_IF5:.*]], label %[[PRED_STORE_CONTINUE6]]
; CHECK:       [[PRED_STORE_IF5]]:
; CHECK-NEXT:    [[TMP11:%.*]] = add i32 [[INDEX]], 3
; CHECK-NEXT:    [[TMP12:%.*]] = getelementptr i32, ptr [[DST]], i32 [[TMP11]]
; CHECK-NEXT:    store i32 [[TMP11]], ptr [[TMP12]], align 8
; CHECK-NEXT:    br label %[[PRED_STORE_CONTINUE6]]
; CHECK:       [[PRED_STORE_CONTINUE6]]:
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <4 x i8> [[VEC_IND]], splat (i8 4)
; CHECK-NEXT:    [[TMP13:%.*]] = icmp eq i32 [[INDEX_NEXT]], 12
; CHECK-NEXT:    br i1 [[TMP13]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP11:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  %cmp1 = icmp ne i8 0, 0
  %zext1 = zext i1 false to i32
  %trunc1 = trunc i64 0 to i16
  %cmp2 = icmp sgt i16 0, 0
  %sel1 = select i1 false, i16 0, i16 0
  %zext2 = zext i16 0 to i32
  %lshr1 = lshr i32 0, 0
  %trunc2 = trunc i32 0 to i16
  %xor1 = xor i16 0, 0
  %zext3 = zext i16 0 to i32
  %ld = load i32, ptr %src, align 4
  %cmp3 = icmp sgt i32 0, 0
  %lshr2 = lshr i32 0, 0
  %cmp4 = icmp ne i32 0, 0
  %sel2 = select i1 false, i1 false, i1 false
  %shl1 = shl i32 0, 0
  %sel3 = select i1 false, i32 0, i32 0
  %sext1 = sext i8 0 to i32
  %cmp5 = icmp eq i32 0, 0
  %zext4 = zext i32 0 to i64
  %sub1 = sub i64 0, 0
  %sext2 = sext i8 0 to i64
  %and1 = and i64 0, 0
  %cmp6 = icmp ne i64 0, 0
  %zext5 = zext i1 false to i16
  %cmp7 = icmp eq i16 0, 0
  br label %loop

loop:
  %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr i32, ptr %dst, i32 %iv
  store i32 %iv, ptr %gep, align 8
  %iv.next = add i32 %iv, 1
  %ec = icmp eq i32 %iv, 8
  br i1 %ec, label %exit, label %loop

exit:
  ret void
}