xref: /llvm-project/llvm/test/Transforms/PhaseOrdering/X86/pr94546.ll (revision 611401c11594871aa5c7692cd17a7f12b6fbe660)
1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2; RUN: opt -mtriple=x86_64-- -mcpu=x86-64    -O3                   -S < %s  | FileCheck %s --check-prefixes=SSE2
3; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v2 -O3                   -S < %s  | FileCheck %s --check-prefixes=SSE4
4; RUN: opt -mtriple=x86_64-- -mcpu=btver2    -O3                   -S < %s  | FileCheck %s --check-prefixes=AVX
5; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v3 -O3                   -S < %s  | FileCheck %s --check-prefixes=AVX
6; RUN: opt -mtriple=x86_64-- -mcpu=x86-64    -passes="default<O3>" -S < %s  | FileCheck %s --check-prefixes=SSE2
7; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v2 -passes="default<O3>" -S < %s  | FileCheck %s --check-prefixes=SSE4
8; RUN: opt -mtriple=x86_64-- -mcpu=btver2    -passes="default<O3>" -S < %s  | FileCheck %s --check-prefixes=AVX
9; RUN: opt -mtriple=x86_64-- -mcpu=x86-64-v3 -passes="default<O3>" -S < %s  | FileCheck %s --check-prefixes=AVX
10
11define <4 x double> @PR94546(<4 x double> %a, <4 x double> %b) {
12; SSE2-LABEL: @PR94546(
13; SSE2-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A:%.*]], <4 x double> [[B:%.*]], <2 x i32> <i32 0, i32 6>
14; SSE2-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[A]], <4 x double> [[B]], <2 x i32> <i32 1, i32 7>
15; SSE2-NEXT:    [[TMP3:%.*]] = fadd <2 x double> [[TMP1]], [[TMP2]]
16; SSE2-NEXT:    [[TMP4:%.*]] = shufflevector <2 x double> [[TMP3]], <2 x double> poison, <4 x i32> <i32 0, i32 poison, i32 poison, i32 1>
17; SSE2-NEXT:    ret <4 x double> [[TMP4]]
18;
19; SSE4-LABEL: @PR94546(
20; SSE4-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A:%.*]], <4 x double> [[B:%.*]], <4 x i32> <i32 0, i32 poison, i32 poison, i32 6>
21; SSE4-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[A]], <4 x double> [[B]], <4 x i32> <i32 1, i32 poison, i32 poison, i32 7>
22; SSE4-NEXT:    [[TMP3:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
23; SSE4-NEXT:    ret <4 x double> [[TMP3]]
24;
25; AVX-LABEL: @PR94546(
26; AVX-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A:%.*]], <4 x double> [[B:%.*]], <4 x i32> <i32 0, i32 poison, i32 poison, i32 6>
27; AVX-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[A]], <4 x double> [[B]], <4 x i32> <i32 1, i32 poison, i32 poison, i32 7>
28; AVX-NEXT:    [[TMP3:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
29; AVX-NEXT:    ret <4 x double> [[TMP3]]
30;
31  %vecext = extractelement <4 x double> %a, i32 0
32  %vecext1 = extractelement <4 x double> %a, i32 1
33  %add = fadd double %vecext, %vecext1
34  %vecinit = insertelement <4 x double> poison, double %add, i32 0
35  %vecext2 = extractelement <4 x double> %a, i32 2
36  %vecext3 = extractelement <4 x double> %a, i32 3
37  %add4 = fadd double %vecext2, %vecext3
38  %vecinit5 = insertelement <4 x double> %vecinit, double %add4, i32 1
39  %vecext6 = extractelement <4 x double> %b, i32 0
40  %vecext7 = extractelement <4 x double> %b, i32 1
41  %add8 = fadd double %vecext6, %vecext7
42  %vecinit9 = insertelement <4 x double> %vecinit5, double %add8, i32 2
43  %vecext10 = extractelement <4 x double> %b, i32 2
44  %vecext11 = extractelement <4 x double> %b, i32 3
45  %add12 = fadd double %vecext10, %vecext11
46  %vecinit13 = insertelement <4 x double> %vecinit9, double %add12, i32 3
47  %shuffle = shufflevector <4 x double> %vecinit13, <4 x double> %a, <4 x i32> <i32 0, i32 poison, i32 poison, i32 3>
48  ret <4 x double> %shuffle
49}
50