TheRock/compiler/amd-llvm/llvm/test/Transforms/PhaseOrdering/X86/vdiv-nounroll.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 6
; RUN: opt < %s -O3 -S | FileCheck %s

target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx11.0.0"

; We expect the loop to be vectorized even if it
; is not unrolled.
; Verify that fdiv is not sunk back into the loop
; after getting hoisted - based on this C source:
;
;  void vdiv(ptr a, float b) {
;      for (int i = 0; i != 1024; ++i)
;        a[i] /= b;
;  }

define void @vdiv(ptr %a, float %b) {
; CHECK-LABEL: define void @vdiv(
; CHECK-SAME: ptr captures(none) [[A:%.*]], float [[B:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x float> poison, float [[B]], i64 0
; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x float> [[BROADCAST_SPLATINSERT]], <4 x float> poison, <4 x i32> zeroinitializer
; CHECK-NEXT:    [[TMP0:%.*]] = fdiv fast <4 x float> splat (float 1.000000e+00), [[BROADCAST_SPLAT]]
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr inbounds nuw [4 x i8], ptr [[A]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP1]], align 4, !tbaa [[FLOAT_TBAA3:![0-9]+]]
; CHECK-NEXT:    [[TMP3:%.*]] = fmul fast <4 x float> [[WIDE_LOAD]], [[TMP0]]
; CHECK-NEXT:    store <4 x float> [[TMP3]], ptr [[TMP1]], align 4, !tbaa [[FLOAT_TBAA3]]
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], 1024
; CHECK-NEXT:    br i1 [[TMP5]], label %[[FOR_COND_CLEANUP:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
; CHECK:       [[FOR_COND_CLEANUP]]:
; CHECK-NEXT:    ret void
;
entry:
  %a.addr = alloca ptr, align 8
  %b.addr = alloca float, align 4
  %i = alloca i32, align 4
  store ptr %a, ptr %a.addr, align 8, !tbaa !3
  store float %b, ptr %b.addr, align 4, !tbaa !7
  call void @llvm.lifetime.start.p0(ptr %i)
  store i32 0, ptr %i, align 4, !tbaa !9
  br label %for.cond

for.cond:                                         ; preds = %for.inc, %entry
  %0 = load i32, ptr %i, align 4, !tbaa !9
  %cmp = icmp ne i32 %0, 1024
  br i1 %cmp, label %for.body, label %for.cond.cleanup

for.cond.cleanup:                                 ; preds = %for.cond
  call void @llvm.lifetime.end.p0(ptr %i)
  br label %for.end

for.body:                                         ; preds = %for.cond
  %1 = load float, ptr %b.addr, align 4, !tbaa !7
  %2 = load ptr, ptr %a.addr, align 8, !tbaa !3
  %3 = load i32, ptr %i, align 4, !tbaa !9
  %idxprom = sext i32 %3 to i64
  %arrayidx = getelementptr inbounds float, ptr %2, i64 %idxprom
  %4 = load float, ptr %arrayidx, align 4, !tbaa !7
  %div = fdiv fast float %4, %1
  store float %div, ptr %arrayidx, align 4, !tbaa !7
  br label %for.inc

for.inc:                                          ; preds = %for.body
  %5 = load i32, ptr %i, align 4, !tbaa !9
  %inc = add nsw i32 %5, 1
  store i32 %inc, ptr %i, align 4, !tbaa !9
  br label %for.cond, !llvm.loop !11

for.end:                                          ; preds = %for.cond.cleanup
  ret void
}

declare void @llvm.lifetime.start.p0(ptr nocapture)
declare void @llvm.lifetime.end.p0(ptr nocapture)

!llvm.module.flags = !{!0, !1}
!llvm.ident = !{!2}

!0 = !{i32 1, !"wchar_size", i32 4}
!1 = !{i32 7, !"PIC Level", i32 2}
!2 = !{!"clang version 13.0.0 (https://github.com/llvm/llvm-project.git 437fb428178716db014f75701f3cdea6dea63f59)"}
!3 = !{!4, !4, i64 0}
!4 = !{!"any pointer", !5, i64 0}
!5 = !{!"omnipotent char", !6, i64 0}
!6 = !{!"Simple C/C++ TBAA"}
!7 = !{!8, !8, i64 0}
!8 = !{!"float", !5, i64 0}
!9 = !{!10, !10, i64 0}
!10 = !{!"int", !5, i64 0}
!11 = distinct !{!11, !12, !13}
!12 = !{!"llvm.loop.mustprogress"}
!13 = !{!"llvm.loop.unroll.disable"}
;.
; CHECK: [[FLOAT_TBAA3]] = !{[[META4:![0-9]+]], [[META4]], i64 0}
; CHECK: [[META4]] = !{!"float", [[META5:![0-9]+]], i64 0}
; CHECK: [[META5]] = !{!"omnipotent char", [[META6:![0-9]+]], i64 0}
; CHECK: [[META6]] = !{!"Simple C/C++ TBAA"}
; CHECK: [[LOOP7]] = distinct !{[[LOOP7]], [[META8:![0-9]+]], [[META9:![0-9]+]], [[META10:![0-9]+]], [[META11:![0-9]+]]}
; CHECK: [[META8]] = !{!"llvm.loop.mustprogress"}
; CHECK: [[META9]] = !{!"llvm.loop.unroll.disable"}
; CHECK: [[META10]] = !{!"llvm.loop.isvectorized", i32 1}
; CHECK: [[META11]] = !{!"llvm.loop.unroll.runtime.disable"}
;.