llvm/test/Transforms/SLPVectorizer/X86/split-node-no-reorder-copy.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt -S --passes=slp-vectorizer -mtriple=x86_64-sie-ps5 -mcpu=znver2 < %s | FileCheck %s

define i1 @test(ptr %0, ptr %1, <2 x float> %2, <2 x float> %3, <2 x float> %4) {
; CHECK-LABEL: define i1 @test(
; CHECK-SAME: ptr [[TMP0:%.*]], ptr [[TMP1:%.*]], <2 x float> [[TMP2:%.*]], <2 x float> [[TMP3:%.*]], <2 x float> [[TMP4:%.*]]) #[[ATTR0:[0-9]+]] {
; CHECK-NEXT:    [[TMP6:%.*]] = load float, ptr [[TMP1]], align 4
; CHECK-NEXT:    [[TMP7:%.*]] = load float, ptr [[TMP0]], align 4
; CHECK-NEXT:    [[TMP8:%.*]] = load float, ptr [[TMP0]], align 4
; CHECK-NEXT:    [[TMP9:%.*]] = extractelement <2 x float> [[TMP4]], i64 0
; CHECK-NEXT:    [[TMP10:%.*]] = insertelement <8 x float> poison, float [[TMP6]], i32 0
; CHECK-NEXT:    [[TMP11:%.*]] = insertelement <8 x float> [[TMP10]], float [[TMP7]], i32 1
; CHECK-NEXT:    [[TMP12:%.*]] = insertelement <8 x float> [[TMP11]], float [[TMP8]], i32 6
; CHECK-NEXT:    [[TMP13:%.*]] = shufflevector <8 x float> [[TMP12]], <8 x float> poison, <8 x i32> <i32 0, i32 1, i32 1, i32 1, i32 1, i32 1, i32 6, i32 6>
; CHECK-NEXT:    [[TMP14:%.*]] = shufflevector <2 x float> [[TMP2]], <2 x float> [[TMP3]], <8 x i32> <i32 0, i32 0, i32 2, i32 0, i32 0, i32 0, i32 0, i32 poison>
; CHECK-NEXT:    [[TMP15:%.*]] = insertelement <8 x float> [[TMP14]], float [[TMP9]], i32 7
; CHECK-NEXT:    [[TMP16:%.*]] = shufflevector <8 x float> [[TMP13]], <8 x float> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison, i32 poison>
; CHECK-NEXT:    [[TMP17:%.*]] = call <16 x float> @llvm.vector.insert.v16f32.v8f32(<16 x float> [[TMP16]], <8 x float> [[TMP15]], i64 8)
; CHECK-NEXT:    [[TMP18:%.*]] = call <16 x float> @llvm.vector.insert.v16f32.v8f32(<16 x float> [[TMP16]], <8 x float> [[TMP15]], i64 8)
; CHECK-NEXT:    [[TMP20:%.*]] = fmul <16 x float> [[TMP18]], [[TMP17]]
; CHECK-NEXT:    [[TMP21:%.*]] = call reassoc nsz float @llvm.vector.reduce.fadd.v16f32(float 0.000000e+00, <16 x float> [[TMP20]])
; CHECK-NEXT:    [[TMP22:%.*]] = call float @foo(float [[TMP21]])
; CHECK-NEXT:    ret i1 false
;
  %6 = load float, ptr %1, align 4
  %7 = load float, ptr %0, align 4
  %8 = fmul float %6, %6
  %9 = fmul float %7, %7
  %10 = fmul float %7, %7
  %11 = fmul float %7, %7
  %12 = fmul float %7, %7
  %13 = fmul float %7, %7
  %14 = load float, ptr %0, align 4
  %15 = fmul float %14, %14
  %16 = fmul float %14, %14
  %17 = extractelement <2 x float> %2, i64 0
  %18 = extractelement <2 x float> %2, i64 0
  %19 = fmul float %17, %17
  %20 = fmul float %18, %18
  %21 = extractelement <2 x float> %3, i64 0
  %22 = extractelement <2 x float> %2, i64 0
  %23 = fmul float %21, %21
  %24 = fmul float %22, %22
  %25 = extractelement <2 x float> %2, i64 0
  %26 = extractelement <2 x float> %2, i64 0
  %27 = fmul float %25, %25
  %28 = fmul float %26, %26
  %29 = extractelement <2 x float> %2, i64 0
  %30 = extractelement <2 x float> %4, i64 0
  %31 = fmul float %29, %29
  %32 = fmul float %30, %30
  %33 = fadd reassoc nsz float %8, %9
  %34 = fadd reassoc nsz float %33, %10
  %35 = fadd reassoc nsz float %34, %11
  %36 = fadd reassoc nsz float %35, %12
  %37 = fadd reassoc nsz float %36, %13
  %38 = fadd reassoc nsz float %37, %15
  %39 = fadd reassoc nsz float %38, %16
  %40 = fadd reassoc nsz float %39, %19
  %41 = fadd reassoc nsz float %40, %20
  %42 = fadd reassoc nsz float %41, %23
  %43 = fadd reassoc nsz float %42, %24
  %44 = fadd reassoc nsz float %43, %27
  %45 = fadd reassoc nsz float %44, %28
  %46 = fadd reassoc nsz float %45, %31
  %47 = fadd reassoc nsz float %46, %32
  %48 = call float @foo(float %47)
  ret i1 false
}

declare float @foo(float)