; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2
; RUN: opt -S < %s -passes=loop-vectorize -force-vector-interleave=1 2>&1 | FileCheck %s
target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
target triple = "aarch64"
define void @add_a(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i32 %len) #0 {
; CHECK-LABEL: define void @add_a
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[LEN]], 0
; CHECK-NEXT: br i1 [[CMP8]], label [[ITER_CHECK:%.*]], label [[FOR_COND_CLEANUP:%.*]]
; CHECK: iter.check:
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[LEN]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 8
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
; CHECK: vector.main.loop.iter.check:
; CHECK-NEXT: [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[TMP0]], 16
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 16
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP1:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr inbounds i8, ptr [[TMP2]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <16 x i8>, ptr [[TMP3]], align 1
; CHECK-NEXT: [[TMP4:%.*]] = add <16 x i8> [[WIDE_LOAD]], <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>
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds i8, ptr [[TMP5]], i32 0
; CHECK-NEXT: store <16 x i8> [[TMP4]], ptr [[TMP6]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEXT: [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
; CHECK: vec.epilog.iter.check:
; CHECK-NEXT: [[N_VEC_REMAINING:%.*]] = sub i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 8
; CHECK-NEXT: br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT: [[N_MOD_VF2:%.*]] = urem i64 [[TMP0]], 8
; CHECK-NEXT: [[N_VEC3:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF2]]
; CHECK-NEXT: br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
; CHECK: vec.epilog.vector.body:
; CHECK-NEXT: [[INDEX5:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT7:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP8:%.*]] = add i64 [[INDEX5]], 0
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP8]]
; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds i8, ptr [[TMP9]], i32 0
; CHECK-NEXT: [[WIDE_LOAD6:%.*]] = load <8 x i8>, ptr [[TMP10]], align 1
; CHECK-NEXT: [[TMP11:%.*]] = add <8 x i8> [[WIDE_LOAD6]], <i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2>
; CHECK-NEXT: [[TMP12:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP8]]
; CHECK-NEXT: [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[TMP12]], i32 0
; CHECK-NEXT: store <8 x i8> [[TMP11]], ptr [[TMP13]], align 1
; CHECK-NEXT: [[INDEX_NEXT7]] = add nuw i64 [[INDEX5]], 8
; CHECK-NEXT: [[TMP14:%.*]] = icmp eq i64 [[INDEX_NEXT7]], [[N_VEC3]]
; CHECK-NEXT: br i1 [[TMP14]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK: vec.epilog.middle.block:
; CHECK-NEXT: [[CMP_N4:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC3]]
; CHECK-NEXT: br i1 [[CMP_N4]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC3]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup.loopexit:
; CHECK-NEXT: br label [[FOR_COND_CLEANUP]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP15:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
; CHECK-NEXT: [[CONV:%.*]] = zext i8 [[TMP15]] to i32
; CHECK-NEXT: [[ADD:%.*]] = add nuw nsw i32 [[CONV]], 2
; CHECK-NEXT: [[CONV1:%.*]] = trunc i32 [[ADD]] to i8
; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i8 [[CONV1]], ptr [[ARRAYIDX3]], align 1
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
;
entry:
%cmp8 = icmp sgt i32 %len, 0
br i1 %cmp8, label %for.body, label %for.cond.cleanup
for.cond.cleanup: ; preds = %for.body, %entry
ret void
for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
%arrayidx = getelementptr inbounds i8, ptr %p, i64 %indvars.iv
%0 = load i8, ptr %arrayidx
%conv = zext i8 %0 to i32
%add = add nuw nsw i32 %conv, 2
%conv1 = trunc i32 %add to i8
%arrayidx3 = getelementptr inbounds i8, ptr %q, i64 %indvars.iv
store i8 %conv1, ptr %arrayidx3
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
; Ensure that we preserve nuw/nsw if we're not shrinking the values we're
; working with.
define void @add_a1(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i32 %len) #0 {
; CHECK-LABEL: define void @add_a1
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[LEN]], 0
; CHECK-NEXT: br i1 [[CMP8]], label [[ITER_CHECK:%.*]], label [[FOR_COND_CLEANUP:%.*]]
; CHECK: iter.check:
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[LEN]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 8
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
; CHECK: vector.main.loop.iter.check:
; CHECK-NEXT: [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[TMP0]], 16
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 16
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP1:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr inbounds i8, ptr [[TMP2]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <16 x i8>, ptr [[TMP3]], align 1
; CHECK-NEXT: [[TMP4:%.*]] = add nuw nsw <16 x i8> [[WIDE_LOAD]], <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>
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds i8, ptr [[TMP5]], i32 0
; CHECK-NEXT: store <16 x i8> [[TMP4]], ptr [[TMP6]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEXT: [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
; CHECK: vec.epilog.iter.check:
; CHECK-NEXT: [[N_VEC_REMAINING:%.*]] = sub i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 8
; CHECK-NEXT: br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT: [[N_MOD_VF2:%.*]] = urem i64 [[TMP0]], 8
; CHECK-NEXT: [[N_VEC3:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF2]]
; CHECK-NEXT: br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
; CHECK: vec.epilog.vector.body:
; CHECK-NEXT: [[INDEX5:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT7:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP8:%.*]] = add i64 [[INDEX5]], 0
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP8]]
; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds i8, ptr [[TMP9]], i32 0
; CHECK-NEXT: [[WIDE_LOAD6:%.*]] = load <8 x i8>, ptr [[TMP10]], align 1
; CHECK-NEXT: [[TMP11:%.*]] = add nuw nsw <8 x i8> [[WIDE_LOAD6]], <i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2, i8 2>
; CHECK-NEXT: [[TMP12:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP8]]
; CHECK-NEXT: [[TMP13:%.*]] = getelementptr inbounds i8, ptr [[TMP12]], i32 0
; CHECK-NEXT: store <8 x i8> [[TMP11]], ptr [[TMP13]], align 1
; CHECK-NEXT: [[INDEX_NEXT7]] = add nuw i64 [[INDEX5]], 8
; CHECK-NEXT: [[TMP14:%.*]] = icmp eq i64 [[INDEX_NEXT7]], [[N_VEC3]]
; CHECK-NEXT: br i1 [[TMP14]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; CHECK: vec.epilog.middle.block:
; CHECK-NEXT: [[CMP_N4:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC3]]
; CHECK-NEXT: br i1 [[CMP_N4]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC3]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup.loopexit:
; CHECK-NEXT: br label [[FOR_COND_CLEANUP]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP15:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
; CHECK-NEXT: [[ADD:%.*]] = add nuw nsw i8 [[TMP15]], 2
; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i8 [[ADD]], ptr [[ARRAYIDX3]], align 1
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
;
entry:
%cmp8 = icmp sgt i32 %len, 0
br i1 %cmp8, label %for.body, label %for.cond.cleanup
for.cond.cleanup: ; preds = %for.body, %entry
ret void
for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
%arrayidx = getelementptr inbounds i8, ptr %p, i64 %indvars.iv
%0 = load i8, ptr %arrayidx
%add = add nuw nsw i8 %0, 2
%arrayidx3 = getelementptr inbounds i8, ptr %q, i64 %indvars.iv
store i8 %add, ptr %arrayidx3
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
define void @add_b(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i32 %len) #0 {
; CHECK-LABEL: define void @add_b
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP9:%.*]] = icmp sgt i32 [[LEN]], 0
; CHECK-NEXT: br i1 [[CMP9]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_COND_CLEANUP:%.*]]
; CHECK: for.body.preheader:
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[LEN]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 8
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 8
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP1:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i16, ptr [[P]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr inbounds i16, ptr [[TMP2]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <8 x i16>, ptr [[TMP3]], align 2
; CHECK-NEXT: [[TMP4:%.*]] = add <8 x i16> [[WIDE_LOAD]], <i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2>
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i16, ptr [[Q]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds i16, ptr [[TMP5]], i32 0
; CHECK-NEXT: store <8 x i16> [[TMP4]], ptr [[TMP6]], align 2
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
; CHECK-NEXT: [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup.loopexit:
; CHECK-NEXT: br label [[FOR_COND_CLEANUP]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i16, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP8:%.*]] = load i16, ptr [[ARRAYIDX]], align 2
; CHECK-NEXT: [[CONV8:%.*]] = zext i16 [[TMP8]] to i32
; CHECK-NEXT: [[ADD:%.*]] = add nuw nsw i32 [[CONV8]], 2
; CHECK-NEXT: [[CONV1:%.*]] = trunc i32 [[ADD]] to i16
; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i16, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i16 [[CONV1]], ptr [[ARRAYIDX3]], align 2
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP9:![0-9]+]]
;
entry:
%cmp9 = icmp sgt i32 %len, 0
br i1 %cmp9, label %for.body, label %for.cond.cleanup
for.cond.cleanup: ; preds = %for.body, %entry
ret void
for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
%arrayidx = getelementptr inbounds i16, ptr %p, i64 %indvars.iv
%0 = load i16, ptr %arrayidx
%conv8 = zext i16 %0 to i32
%add = add nuw nsw i32 %conv8, 2
%conv1 = trunc i32 %add to i16
%arrayidx3 = getelementptr inbounds i16, ptr %q, i64 %indvars.iv
store i16 %conv1, ptr %arrayidx3
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
define void @add_c(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i32 %len) #0 {
; CHECK-LABEL: define void @add_c
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[LEN]], 0
; CHECK-NEXT: br i1 [[CMP8]], label [[ITER_CHECK:%.*]], label [[FOR_COND_CLEANUP:%.*]]
; CHECK: iter.check:
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[LEN]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 8
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
; CHECK: vector.main.loop.iter.check:
; CHECK-NEXT: [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[TMP0]], 16
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 16
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP1:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr inbounds i8, ptr [[TMP2]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <16 x i8>, ptr [[TMP3]], align 1
; CHECK-NEXT: [[TMP4:%.*]] = zext <16 x i8> [[WIDE_LOAD]] to <16 x i16>
; CHECK-NEXT: [[TMP5:%.*]] = add <16 x i16> [[TMP4]], <i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2>
; CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds i16, ptr [[Q]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP7:%.*]] = getelementptr inbounds i16, ptr [[TMP6]], i32 0
; CHECK-NEXT: store <16 x i16> [[TMP5]], ptr [[TMP7]], align 2
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEXT: [[TMP8:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
; CHECK: vec.epilog.iter.check:
; CHECK-NEXT: [[N_VEC_REMAINING:%.*]] = sub i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 8
; CHECK-NEXT: br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT: [[N_MOD_VF2:%.*]] = urem i64 [[TMP0]], 8
; CHECK-NEXT: [[N_VEC3:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF2]]
; CHECK-NEXT: br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
; CHECK: vec.epilog.vector.body:
; CHECK-NEXT: [[INDEX5:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT7:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP9:%.*]] = add i64 [[INDEX5]], 0
; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP9]]
; CHECK-NEXT: [[TMP11:%.*]] = getelementptr inbounds i8, ptr [[TMP10]], i32 0
; CHECK-NEXT: [[WIDE_LOAD6:%.*]] = load <8 x i8>, ptr [[TMP11]], align 1
; CHECK-NEXT: [[TMP12:%.*]] = zext <8 x i8> [[WIDE_LOAD6]] to <8 x i16>
; CHECK-NEXT: [[TMP13:%.*]] = add <8 x i16> [[TMP12]], <i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2, i16 2>
; CHECK-NEXT: [[TMP14:%.*]] = getelementptr inbounds i16, ptr [[Q]], i64 [[TMP9]]
; CHECK-NEXT: [[TMP15:%.*]] = getelementptr inbounds i16, ptr [[TMP14]], i32 0
; CHECK-NEXT: store <8 x i16> [[TMP13]], ptr [[TMP15]], align 2
; CHECK-NEXT: [[INDEX_NEXT7]] = add nuw i64 [[INDEX5]], 8
; CHECK-NEXT: [[TMP16:%.*]] = icmp eq i64 [[INDEX_NEXT7]], [[N_VEC3]]
; CHECK-NEXT: br i1 [[TMP16]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP11:![0-9]+]]
; CHECK: vec.epilog.middle.block:
; CHECK-NEXT: [[CMP_N4:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC3]]
; CHECK-NEXT: br i1 [[CMP_N4]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC3]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup.loopexit:
; CHECK-NEXT: br label [[FOR_COND_CLEANUP]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP17:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
; CHECK-NEXT: [[CONV:%.*]] = zext i8 [[TMP17]] to i32
; CHECK-NEXT: [[ADD:%.*]] = add nuw nsw i32 [[CONV]], 2
; CHECK-NEXT: [[CONV1:%.*]] = trunc i32 [[ADD]] to i16
; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i16, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i16 [[CONV1]], ptr [[ARRAYIDX3]], align 2
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
;
entry:
%cmp8 = icmp sgt i32 %len, 0
br i1 %cmp8, label %for.body, label %for.cond.cleanup
for.cond.cleanup: ; preds = %for.body, %entry
ret void
for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
%arrayidx = getelementptr inbounds i8, ptr %p, i64 %indvars.iv
%0 = load i8, ptr %arrayidx
%conv = zext i8 %0 to i32
%add = add nuw nsw i32 %conv, 2
%conv1 = trunc i32 %add to i16
%arrayidx3 = getelementptr inbounds i16, ptr %q, i64 %indvars.iv
store i16 %conv1, ptr %arrayidx3
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
define void @add_d(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i32 %len) #0 {
; CHECK-LABEL: define void @add_d
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP7:%.*]] = icmp sgt i32 [[LEN]], 0
; CHECK-NEXT: br i1 [[CMP7]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_COND_CLEANUP:%.*]]
; CHECK: for.body.preheader:
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[LEN]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 8
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 8
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP1:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i16, ptr [[P]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr inbounds i16, ptr [[TMP2]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <8 x i16>, ptr [[TMP3]], align 2
; CHECK-NEXT: [[TMP4:%.*]] = sext <8 x i16> [[WIDE_LOAD]] to <8 x i32>
; CHECK-NEXT: [[TMP5:%.*]] = add nsw <8 x i32> [[TMP4]], <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
; CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds i32, ptr [[Q]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP7:%.*]] = getelementptr inbounds i32, ptr [[TMP6]], i32 0
; CHECK-NEXT: store <8 x i32> [[TMP5]], ptr [[TMP7]], align 4
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
; CHECK-NEXT: [[TMP8:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP13:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup.loopexit:
; CHECK-NEXT: br label [[FOR_COND_CLEANUP]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i16, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP9:%.*]] = load i16, ptr [[ARRAYIDX]], align 2
; CHECK-NEXT: [[CONV:%.*]] = sext i16 [[TMP9]] to i32
; CHECK-NEXT: [[ADD:%.*]] = add nsw i32 [[CONV]], 2
; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i32 [[ADD]], ptr [[ARRAYIDX2]], align 4
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP14:![0-9]+]]
;
entry:
%cmp7 = icmp sgt i32 %len, 0
br i1 %cmp7, label %for.body, label %for.cond.cleanup
for.cond.cleanup: ; preds = %for.body, %entry
ret void
for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
%arrayidx = getelementptr inbounds i16, ptr %p, i64 %indvars.iv
%0 = load i16, ptr %arrayidx
%conv = sext i16 %0 to i32
%add = add nsw i32 %conv, 2
%arrayidx2 = getelementptr inbounds i32, ptr %q, i64 %indvars.iv
store i32 %add, ptr %arrayidx2
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
define void @add_e(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i8 %arg1, i8 %arg2, i32 %len) #0 {
; CHECK-LABEL: define void @add_e
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i8 [[ARG1:%.*]], i8 [[ARG2:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP_32:%.*]] = icmp sgt i32 [[LEN]], 0
; CHECK-NEXT: br i1 [[CMP_32]], label [[ITER_CHECK:%.*]], label [[FOR_COND_CLEANUP:%.*]]
; CHECK: iter.check:
; CHECK-NEXT: [[CONV11:%.*]] = zext i8 [[ARG2]] to i32
; CHECK-NEXT: [[CONV13:%.*]] = zext i8 [[ARG1]] to i32
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[LEN]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 8
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
; CHECK: vector.main.loop.iter.check:
; CHECK-NEXT: [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[TMP0]], 16
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 16
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <16 x i32> poison, i32 [[CONV13]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <16 x i32> [[BROADCAST_SPLATINSERT]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK-NEXT: [[TMP1:%.*]] = trunc <16 x i32> [[BROADCAST_SPLAT]] to <16 x i8>
; CHECK-NEXT: [[BROADCAST_SPLATINSERT2:%.*]] = insertelement <16 x i32> poison, i32 [[CONV11]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT3:%.*]] = shufflevector <16 x i32> [[BROADCAST_SPLATINSERT2]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK-NEXT: [[TMP2:%.*]] = trunc <16 x i32> [[BROADCAST_SPLAT3]] to <16 x i8>
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP3:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP3]]
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i8, ptr [[TMP4]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <16 x i8>, ptr [[TMP5]], align 1
; CHECK-NEXT: [[TMP6:%.*]] = shl <16 x i8> [[WIDE_LOAD]], <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>
; CHECK-NEXT: [[TMP7:%.*]] = add <16 x i8> [[TMP6]], <i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32>
; CHECK-NEXT: [[TMP8:%.*]] = or <16 x i8> [[WIDE_LOAD]], <i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51>
; CHECK-NEXT: [[TMP9:%.*]] = mul <16 x i8> [[TMP8]], <i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60>
; CHECK-NEXT: [[TMP10:%.*]] = and <16 x i8> [[TMP7]], [[TMP1]]
; CHECK-NEXT: [[TMP11:%.*]] = and <16 x i8> [[TMP9]], <i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4>
; CHECK-NEXT: [[TMP12:%.*]] = xor <16 x i8> [[TMP11]], [[TMP2]]
; CHECK-NEXT: [[TMP13:%.*]] = mul <16 x i8> [[TMP12]], [[TMP10]]
; CHECK-NEXT: [[TMP14:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP3]]
; CHECK-NEXT: [[TMP15:%.*]] = getelementptr inbounds i8, ptr [[TMP14]], i32 0
; CHECK-NEXT: store <16 x i8> [[TMP13]], ptr [[TMP15]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEXT: [[TMP16:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP16]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP15:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
; CHECK: vec.epilog.iter.check:
; CHECK-NEXT: [[N_VEC_REMAINING:%.*]] = sub i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 8
; CHECK-NEXT: br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT: [[N_MOD_VF4:%.*]] = urem i64 [[TMP0]], 8
; CHECK-NEXT: [[N_VEC5:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF4]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT7:%.*]] = insertelement <8 x i32> poison, i32 [[CONV13]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT8:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT7]], <8 x i32> poison, <8 x i32> zeroinitializer
; CHECK-NEXT: [[TMP17:%.*]] = trunc <8 x i32> [[BROADCAST_SPLAT8]] to <8 x i8>
; CHECK-NEXT: [[BROADCAST_SPLATINSERT9:%.*]] = insertelement <8 x i32> poison, i32 [[CONV11]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT10:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT9]], <8 x i32> poison, <8 x i32> zeroinitializer
; CHECK-NEXT: [[TMP18:%.*]] = trunc <8 x i32> [[BROADCAST_SPLAT10]] to <8 x i8>
; CHECK-NEXT: br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
; CHECK: vec.epilog.vector.body:
; CHECK-NEXT: [[INDEX11:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT13:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP19:%.*]] = add i64 [[INDEX11]], 0
; CHECK-NEXT: [[TMP20:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP19]]
; CHECK-NEXT: [[TMP21:%.*]] = getelementptr inbounds i8, ptr [[TMP20]], i32 0
; CHECK-NEXT: [[WIDE_LOAD12:%.*]] = load <8 x i8>, ptr [[TMP21]], align 1
; CHECK-NEXT: [[TMP22:%.*]] = shl <8 x i8> [[WIDE_LOAD12]], <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>
; CHECK-NEXT: [[TMP23:%.*]] = add <8 x i8> [[TMP22]], <i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32>
; CHECK-NEXT: [[TMP24:%.*]] = or <8 x i8> [[WIDE_LOAD12]], <i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51>
; CHECK-NEXT: [[TMP25:%.*]] = mul <8 x i8> [[TMP24]], <i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60>
; CHECK-NEXT: [[TMP26:%.*]] = and <8 x i8> [[TMP23]], [[TMP17]]
; CHECK-NEXT: [[TMP27:%.*]] = and <8 x i8> [[TMP25]], <i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4>
; CHECK-NEXT: [[TMP28:%.*]] = xor <8 x i8> [[TMP27]], [[TMP18]]
; CHECK-NEXT: [[TMP29:%.*]] = mul <8 x i8> [[TMP28]], [[TMP26]]
; CHECK-NEXT: [[TMP30:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP19]]
; CHECK-NEXT: [[TMP31:%.*]] = getelementptr inbounds i8, ptr [[TMP30]], i32 0
; CHECK-NEXT: store <8 x i8> [[TMP29]], ptr [[TMP31]], align 1
; CHECK-NEXT: [[INDEX_NEXT13]] = add nuw i64 [[INDEX11]], 8
; CHECK-NEXT: [[TMP32:%.*]] = icmp eq i64 [[INDEX_NEXT13]], [[N_VEC5]]
; CHECK-NEXT: br i1 [[TMP32]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP16:![0-9]+]]
; CHECK: vec.epilog.middle.block:
; CHECK-NEXT: [[CMP_N6:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC5]]
; CHECK-NEXT: br i1 [[CMP_N6]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC5]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup.loopexit:
; CHECK-NEXT: br label [[FOR_COND_CLEANUP]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP33:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
; CHECK-NEXT: [[CONV:%.*]] = zext i8 [[TMP33]] to i32
; CHECK-NEXT: [[ADD:%.*]] = shl i32 [[CONV]], 4
; CHECK-NEXT: [[CONV2:%.*]] = add nuw nsw i32 [[ADD]], 32
; CHECK-NEXT: [[OR:%.*]] = or i32 [[CONV]], 51
; CHECK-NEXT: [[MUL:%.*]] = mul nuw nsw i32 [[OR]], 60
; CHECK-NEXT: [[AND:%.*]] = and i32 [[CONV2]], [[CONV13]]
; CHECK-NEXT: [[MUL_MASKED:%.*]] = and i32 [[MUL]], 252
; CHECK-NEXT: [[CONV17:%.*]] = xor i32 [[MUL_MASKED]], [[CONV11]]
; CHECK-NEXT: [[MUL18:%.*]] = mul nuw nsw i32 [[CONV17]], [[AND]]
; CHECK-NEXT: [[CONV19:%.*]] = trunc i32 [[MUL18]] to i8
; CHECK-NEXT: [[ARRAYIDX21:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i8 [[CONV19]], ptr [[ARRAYIDX21]], align 1
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP17:![0-9]+]]
;
entry:
%cmp.32 = icmp sgt i32 %len, 0
br i1 %cmp.32, label %for.body.lr.ph, label %for.cond.cleanup
for.body.lr.ph: ; preds = %entry
%conv11 = zext i8 %arg2 to i32
%conv13 = zext i8 %arg1 to i32
br label %for.body
for.cond.cleanup: ; preds = %for.body, %entry
ret void
for.body: ; preds = %for.body, %for.body.lr.ph
%indvars.iv = phi i64 [ 0, %for.body.lr.ph ], [ %indvars.iv.next, %for.body ]
%arrayidx = getelementptr inbounds i8, ptr %p, i64 %indvars.iv
%0 = load i8, ptr %arrayidx
%conv = zext i8 %0 to i32
%add = shl i32 %conv, 4
%conv2 = add nuw nsw i32 %add, 32
%or = or i32 %conv, 51
%mul = mul nuw nsw i32 %or, 60
%and = and i32 %conv2, %conv13
%mul.masked = and i32 %mul, 252
%conv17 = xor i32 %mul.masked, %conv11
%mul18 = mul nuw nsw i32 %conv17, %and
%conv19 = trunc i32 %mul18 to i8
%arrayidx21 = getelementptr inbounds i8, ptr %q, i64 %indvars.iv
store i8 %conv19, ptr %arrayidx21
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
define void @add_f(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i8 %arg1, i8 %arg2, i32 %len) #0 {
; CHECK-LABEL: define void @add_f
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i8 [[ARG1:%.*]], i8 [[ARG2:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP_32:%.*]] = icmp sgt i32 [[LEN]], 0
; CHECK-NEXT: br i1 [[CMP_32]], label [[ITER_CHECK:%.*]], label [[FOR_COND_CLEANUP:%.*]]
; CHECK: iter.check:
; CHECK-NEXT: [[CONV11:%.*]] = zext i8 [[ARG2]] to i32
; CHECK-NEXT: [[CONV13:%.*]] = zext i8 [[ARG1]] to i32
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[LEN]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 8
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
; CHECK: vector.main.loop.iter.check:
; CHECK-NEXT: [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[TMP0]], 16
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 16
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <16 x i32> poison, i32 [[CONV13]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <16 x i32> [[BROADCAST_SPLATINSERT]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK-NEXT: [[TMP1:%.*]] = trunc <16 x i32> [[BROADCAST_SPLAT]] to <16 x i8>
; CHECK-NEXT: [[BROADCAST_SPLATINSERT2:%.*]] = insertelement <16 x i32> poison, i32 [[CONV11]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT3:%.*]] = shufflevector <16 x i32> [[BROADCAST_SPLATINSERT2]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK-NEXT: [[TMP2:%.*]] = trunc <16 x i32> [[BROADCAST_SPLAT3]] to <16 x i8>
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP3:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds i16, ptr [[P]], i64 [[TMP3]]
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i16, ptr [[TMP4]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <16 x i16>, ptr [[TMP5]], align 2
; CHECK-NEXT: [[TMP6:%.*]] = trunc <16 x i16> [[WIDE_LOAD]] to <16 x i8>
; CHECK-NEXT: [[TMP7:%.*]] = shl <16 x i8> [[TMP6]], <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>
; CHECK-NEXT: [[TMP8:%.*]] = add <16 x i8> [[TMP7]], <i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32>
; CHECK-NEXT: [[TMP9:%.*]] = and <16 x i8> [[TMP6]], <i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52>
; CHECK-NEXT: [[TMP10:%.*]] = or <16 x i8> [[TMP9]], <i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51>
; CHECK-NEXT: [[TMP11:%.*]] = mul <16 x i8> [[TMP10]], <i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60>
; CHECK-NEXT: [[TMP12:%.*]] = and <16 x i8> [[TMP8]], [[TMP1]]
; CHECK-NEXT: [[TMP13:%.*]] = and <16 x i8> [[TMP11]], <i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4>
; CHECK-NEXT: [[TMP14:%.*]] = xor <16 x i8> [[TMP13]], [[TMP2]]
; CHECK-NEXT: [[TMP15:%.*]] = mul <16 x i8> [[TMP14]], [[TMP12]]
; CHECK-NEXT: [[TMP16:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP3]]
; CHECK-NEXT: [[TMP17:%.*]] = getelementptr inbounds i8, ptr [[TMP16]], i32 0
; CHECK-NEXT: store <16 x i8> [[TMP15]], ptr [[TMP17]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEXT: [[TMP18:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP18]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP18:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
; CHECK: vec.epilog.iter.check:
; CHECK-NEXT: [[N_VEC_REMAINING:%.*]] = sub i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 8
; CHECK-NEXT: br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT: [[N_MOD_VF4:%.*]] = urem i64 [[TMP0]], 8
; CHECK-NEXT: [[N_VEC5:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF4]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT7:%.*]] = insertelement <8 x i32> poison, i32 [[CONV13]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT8:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT7]], <8 x i32> poison, <8 x i32> zeroinitializer
; CHECK-NEXT: [[TMP19:%.*]] = trunc <8 x i32> [[BROADCAST_SPLAT8]] to <8 x i8>
; CHECK-NEXT: [[BROADCAST_SPLATINSERT9:%.*]] = insertelement <8 x i32> poison, i32 [[CONV11]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT10:%.*]] = shufflevector <8 x i32> [[BROADCAST_SPLATINSERT9]], <8 x i32> poison, <8 x i32> zeroinitializer
; CHECK-NEXT: [[TMP20:%.*]] = trunc <8 x i32> [[BROADCAST_SPLAT10]] to <8 x i8>
; CHECK-NEXT: br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
; CHECK: vec.epilog.vector.body:
; CHECK-NEXT: [[INDEX11:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT13:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP21:%.*]] = add i64 [[INDEX11]], 0
; CHECK-NEXT: [[TMP22:%.*]] = getelementptr inbounds i16, ptr [[P]], i64 [[TMP21]]
; CHECK-NEXT: [[TMP23:%.*]] = getelementptr inbounds i16, ptr [[TMP22]], i32 0
; CHECK-NEXT: [[WIDE_LOAD12:%.*]] = load <8 x i16>, ptr [[TMP23]], align 2
; CHECK-NEXT: [[TMP24:%.*]] = trunc <8 x i16> [[WIDE_LOAD12]] to <8 x i8>
; CHECK-NEXT: [[TMP25:%.*]] = shl <8 x i8> [[TMP24]], <i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4, i8 4>
; CHECK-NEXT: [[TMP26:%.*]] = add <8 x i8> [[TMP25]], <i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32, i8 32>
; CHECK-NEXT: [[TMP27:%.*]] = and <8 x i8> [[TMP24]], <i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52, i8 -52>
; CHECK-NEXT: [[TMP28:%.*]] = or <8 x i8> [[TMP27]], <i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51, i8 51>
; CHECK-NEXT: [[TMP29:%.*]] = mul <8 x i8> [[TMP28]], <i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60, i8 60>
; CHECK-NEXT: [[TMP30:%.*]] = and <8 x i8> [[TMP26]], [[TMP19]]
; CHECK-NEXT: [[TMP31:%.*]] = and <8 x i8> [[TMP29]], <i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4, i8 -4>
; CHECK-NEXT: [[TMP32:%.*]] = xor <8 x i8> [[TMP31]], [[TMP20]]
; CHECK-NEXT: [[TMP33:%.*]] = mul <8 x i8> [[TMP32]], [[TMP30]]
; CHECK-NEXT: [[TMP34:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP21]]
; CHECK-NEXT: [[TMP35:%.*]] = getelementptr inbounds i8, ptr [[TMP34]], i32 0
; CHECK-NEXT: store <8 x i8> [[TMP33]], ptr [[TMP35]], align 1
; CHECK-NEXT: [[INDEX_NEXT13]] = add nuw i64 [[INDEX11]], 8
; CHECK-NEXT: [[TMP36:%.*]] = icmp eq i64 [[INDEX_NEXT13]], [[N_VEC5]]
; CHECK-NEXT: br i1 [[TMP36]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP19:![0-9]+]]
; CHECK: vec.epilog.middle.block:
; CHECK-NEXT: [[CMP_N6:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC5]]
; CHECK-NEXT: br i1 [[CMP_N6]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC5]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup.loopexit:
; CHECK-NEXT: br label [[FOR_COND_CLEANUP]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i16, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP37:%.*]] = load i16, ptr [[ARRAYIDX]], align 2
; CHECK-NEXT: [[CONV:%.*]] = sext i16 [[TMP37]] to i32
; CHECK-NEXT: [[ADD:%.*]] = shl i32 [[CONV]], 4
; CHECK-NEXT: [[CONV2:%.*]] = add nsw i32 [[ADD]], 32
; CHECK-NEXT: [[OR:%.*]] = and i32 [[CONV]], 204
; CHECK-NEXT: [[CONV8:%.*]] = or i32 [[OR]], 51
; CHECK-NEXT: [[MUL:%.*]] = mul nuw nsw i32 [[CONV8]], 60
; CHECK-NEXT: [[AND:%.*]] = and i32 [[CONV2]], [[CONV13]]
; CHECK-NEXT: [[MUL_MASKED:%.*]] = and i32 [[MUL]], 252
; CHECK-NEXT: [[CONV17:%.*]] = xor i32 [[MUL_MASKED]], [[CONV11]]
; CHECK-NEXT: [[MUL18:%.*]] = mul nuw nsw i32 [[CONV17]], [[AND]]
; CHECK-NEXT: [[CONV19:%.*]] = trunc i32 [[MUL18]] to i8
; CHECK-NEXT: [[ARRAYIDX21:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i8 [[CONV19]], ptr [[ARRAYIDX21]], align 1
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP20:![0-9]+]]
;
entry:
%cmp.32 = icmp sgt i32 %len, 0
br i1 %cmp.32, label %for.body.lr.ph, label %for.cond.cleanup
for.body.lr.ph: ; preds = %entry
%conv11 = zext i8 %arg2 to i32
%conv13 = zext i8 %arg1 to i32
br label %for.body
for.cond.cleanup: ; preds = %for.body, %entry
ret void
for.body: ; preds = %for.body, %for.body.lr.ph
%indvars.iv = phi i64 [ 0, %for.body.lr.ph ], [ %indvars.iv.next, %for.body ]
%arrayidx = getelementptr inbounds i16, ptr %p, i64 %indvars.iv
%0 = load i16, ptr %arrayidx
%conv = sext i16 %0 to i32
%add = shl i32 %conv, 4
%conv2 = add nsw i32 %add, 32
%or = and i32 %conv, 204
%conv8 = or i32 %or, 51
%mul = mul nuw nsw i32 %conv8, 60
%and = and i32 %conv2, %conv13
%mul.masked = and i32 %mul, 252
%conv17 = xor i32 %mul.masked, %conv11
%mul18 = mul nuw nsw i32 %conv17, %and
%conv19 = trunc i32 %mul18 to i8
%arrayidx21 = getelementptr inbounds i8, ptr %q, i64 %indvars.iv
store i8 %conv19, ptr %arrayidx21
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
define void @add_phifail(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i32 %len) #0 {
; CHECK-LABEL: define void @add_phifail
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP8:%.*]] = icmp sgt i32 [[LEN]], 0
; CHECK-NEXT: br i1 [[CMP8]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_COND_CLEANUP:%.*]]
; CHECK: for.body.preheader:
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[LEN]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 16
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 16
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[VECTOR_RECUR:%.*]] = phi <16 x i32> [ <i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 0>, [[VECTOR_PH]] ], [ [[TMP4:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP1:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr inbounds i8, ptr [[TMP2]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <16 x i8>, ptr [[TMP3]], align 1
; CHECK-NEXT: [[TMP4]] = zext <16 x i8> [[WIDE_LOAD]] to <16 x i32>
; CHECK-NEXT: [[TMP6:%.*]] = add nuw nsw <16 x i32> [[TMP4]], <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
; CHECK-NEXT: [[TMP7:%.*]] = trunc <16 x i32> [[TMP6]] to <16 x i8>
; CHECK-NEXT: [[TMP8:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds i8, ptr [[TMP8]], i32 0
; CHECK-NEXT: store <16 x i8> [[TMP7]], ptr [[TMP9]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEXT: [[TMP10:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP10]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP21:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <16 x i32> [[TMP4]], i32 15
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP_LOOPEXIT:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: [[SCALAR_RECUR_INIT:%.*]] = phi i32 [ [[VECTOR_RECUR_EXTRACT]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup.loopexit:
; CHECK-NEXT: br label [[FOR_COND_CLEANUP]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: ret void
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[SCALAR_RECUR:%.*]] = phi i32 [ [[CONV:%.*]], [[FOR_BODY]] ], [ [[SCALAR_RECUR_INIT]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP11:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
; CHECK-NEXT: [[CONV]] = zext i8 [[TMP11]] to i32
; CHECK-NEXT: [[ADD:%.*]] = add nuw nsw i32 [[CONV]], 2
; CHECK-NEXT: [[CONV1:%.*]] = trunc i32 [[ADD]] to i8
; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i8 [[CONV1]], ptr [[ARRAYIDX3]], align 1
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP22:![0-9]+]]
;
entry:
%cmp8 = icmp sgt i32 %len, 0
br i1 %cmp8, label %for.body, label %for.cond.cleanup
for.cond.cleanup: ; preds = %for.body, %entry
ret void
for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
%a_phi = phi i32 [ %conv, %for.body ], [ 0, %entry ]
%arrayidx = getelementptr inbounds i8, ptr %p, i64 %indvars.iv
%0 = load i8, ptr %arrayidx
%conv = zext i8 %0 to i32
%add = add nuw nsw i32 %conv, 2
%conv1 = trunc i32 %add to i8
%arrayidx3 = getelementptr inbounds i8, ptr %q, i64 %indvars.iv
store i8 %conv1, ptr %arrayidx3
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
; When we vectorize this loop, we generate correct code
; even when %len exactly divides VF (since we extract from the second last index
; and pass this to the for.cond.cleanup block). Vectorized loop returns
; the correct value a_phi = p[len -2]
define i8 @add_phifail2(ptr noalias nocapture readonly %p, ptr noalias nocapture %q, i32 %len) #0 {
; CHECK-LABEL: define i8 @add_phifail2
; CHECK-SAME: (ptr noalias nocapture readonly [[P:%.*]], ptr noalias nocapture [[Q:%.*]], i32 [[LEN:%.*]]) {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[TMP0:%.*]] = add i32 [[LEN]], -1
; CHECK-NEXT: [[TMP1:%.*]] = zext i32 [[TMP0]] to i64
; CHECK-NEXT: [[TMP2:%.*]] = add nuw nsw i64 [[TMP1]], 1
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP2]], 16
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP2]], 16
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP2]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[VECTOR_RECUR:%.*]] = phi <16 x i32> [ <i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 0>, [[VECTOR_PH]] ], [ [[TMP6:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP3:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[TMP3]]
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds i8, ptr [[TMP4]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <16 x i8>, ptr [[TMP5]], align 1
; CHECK-NEXT: [[TMP6]] = zext <16 x i8> [[WIDE_LOAD]] to <16 x i32>
; CHECK-NEXT: [[TMP8:%.*]] = add nuw nsw <16 x i32> [[TMP6]], <i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2, i32 2>
; CHECK-NEXT: [[TMP9:%.*]] = trunc <16 x i32> [[TMP8]] to <16 x i8>
; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[TMP3]]
; CHECK-NEXT: [[TMP11:%.*]] = getelementptr inbounds i8, ptr [[TMP10]], i32 0
; CHECK-NEXT: store <16 x i8> [[TMP9]], ptr [[TMP11]], align 1
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEXT: [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP23:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <16 x i32> [[TMP6]], i32 15
; CHECK-NEXT: [[VECTOR_RECUR_EXTRACT_FOR_PHI:%.*]] = extractelement <16 x i32> [[TMP6]], i32 14
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP2]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[SCALAR_RECUR_INIT:%.*]] = phi i32 [ [[VECTOR_RECUR_EXTRACT]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.cond.cleanup:
; CHECK-NEXT: [[A_PHI_LCSSA:%.*]] = phi i32 [ [[SCALAR_RECUR:%.*]], [[FOR_BODY]] ], [ [[VECTOR_RECUR_EXTRACT_FOR_PHI]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: [[RET:%.*]] = trunc i32 [[A_PHI_LCSSA]] to i8
; CHECK-NEXT: ret i8 [[RET]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[SCALAR_RECUR]] = phi i32 [ [[CONV:%.*]], [[FOR_BODY]] ], [ [[SCALAR_RECUR_INIT]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i8, ptr [[P]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP13:%.*]] = load i8, ptr [[ARRAYIDX]], align 1
; CHECK-NEXT: [[CONV]] = zext i8 [[TMP13]] to i32
; CHECK-NEXT: [[ADD:%.*]] = add nuw nsw i32 [[CONV]], 2
; CHECK-NEXT: [[CONV1:%.*]] = trunc i32 [[ADD]] to i8
; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i8, ptr [[Q]], i64 [[INDVARS_IV]]
; CHECK-NEXT: store i8 [[CONV1]], ptr [[ARRAYIDX3]], align 1
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[LEN]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND_CLEANUP]], label [[FOR_BODY]], !llvm.loop [[LOOP24:![0-9]+]]
;
entry:
br label %for.body
for.cond.cleanup: ; preds = %for.body, %entry
%ret = trunc i32 %a_phi to i8
ret i8 %ret
for.body: ; preds = %entry, %for.body
%indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %entry ]
%a_phi = phi i32 [ %conv, %for.body ], [ 0, %entry ]
%arrayidx = getelementptr inbounds i8, ptr %p, i64 %indvars.iv
%0 = load i8, ptr %arrayidx
%conv = zext i8 %0 to i32
%add = add nuw nsw i32 %conv, 2
%conv1 = trunc i32 %add to i8
%arrayidx3 = getelementptr inbounds i8, ptr %q, i64 %indvars.iv
store i8 %conv1, ptr %arrayidx3
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %len
br i1 %exitcond, label %for.cond.cleanup, label %for.body
}
Codebase Browser
llvm