File: avx2-vperm.ll

package info (click to toggle)
llvm-toolchain-11 1%3A11.0.1-2
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 995,808 kB
  • sloc: cpp: 4,767,656; ansic: 760,916; asm: 477,436; python: 170,940; objc: 69,804; lisp: 29,914; sh: 23,855; f90: 18,173; pascal: 7,551; perl: 7,471; ml: 5,603; awk: 3,489; makefile: 2,573; xml: 915; cs: 573; fortran: 503; javascript: 452
file content (68 lines) | stat: -rw-r--r-- 2,386 bytes parent folder | download | duplicates (15)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=X32
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=X64

define <8 x i32> @perm_cl_int_8x32(<8 x i32> %A) nounwind readnone {
; X32-LABEL: perm_cl_int_8x32:
; X32:       # %bb.0: # %entry
; X32-NEXT:    vmovaps {{.*#+}} ymm1 = [0,7,2,1,2,7,6,0]
; X32-NEXT:    vpermps %ymm0, %ymm1, %ymm0
; X32-NEXT:    retl
;
; X64-LABEL: perm_cl_int_8x32:
; X64:       # %bb.0: # %entry
; X64-NEXT:    vmovaps {{.*#+}} ymm1 = [0,7,2,1,2,7,6,0]
; X64-NEXT:    vpermps %ymm0, %ymm1, %ymm0
; X64-NEXT:    retq
entry:
  %B = shufflevector <8 x i32> %A, <8 x i32> undef, <8 x i32> <i32 0, i32 7, i32 2, i32 1, i32 2, i32 7, i32 6, i32 0>
  ret <8 x i32> %B
}


define <8 x float> @perm_cl_fp_8x32(<8 x float> %A) nounwind readnone {
; X32-LABEL: perm_cl_fp_8x32:
; X32:       # %bb.0: # %entry
; X32-NEXT:    vmovaps {{.*#+}} ymm1 = <u,7,2,u,4,u,1,6>
; X32-NEXT:    vpermps %ymm0, %ymm1, %ymm0
; X32-NEXT:    retl
;
; X64-LABEL: perm_cl_fp_8x32:
; X64:       # %bb.0: # %entry
; X64-NEXT:    vmovaps {{.*#+}} ymm1 = <u,7,2,u,4,u,1,6>
; X64-NEXT:    vpermps %ymm0, %ymm1, %ymm0
; X64-NEXT:    retq
entry:
  %B = shufflevector <8 x float> %A, <8 x float> undef, <8 x i32> <i32 undef, i32 7, i32 2, i32 undef, i32 4, i32 undef, i32 1, i32 6>
  ret <8 x float> %B
}

define <4 x i64> @perm_cl_int_4x64(<4 x i64> %A) nounwind readnone {
; X32-LABEL: perm_cl_int_4x64:
; X32:       # %bb.0: # %entry
; X32-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,3,2,1]
; X32-NEXT:    retl
;
; X64-LABEL: perm_cl_int_4x64:
; X64:       # %bb.0: # %entry
; X64-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,3,2,1]
; X64-NEXT:    retq
entry:
  %B = shufflevector <4 x i64> %A, <4 x i64> undef, <4 x i32> <i32 0, i32 3, i32 2, i32 1>
  ret <4 x i64> %B
}

define <4 x double> @perm_cl_fp_4x64(<4 x double> %A) nounwind readnone {
; X32-LABEL: perm_cl_fp_4x64:
; X32:       # %bb.0: # %entry
; X32-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,3,2,1]
; X32-NEXT:    retl
;
; X64-LABEL: perm_cl_fp_4x64:
; X64:       # %bb.0: # %entry
; X64-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,3,2,1]
; X64-NEXT:    retq
entry:
  %B = shufflevector <4 x double> %A, <4 x double> undef, <4 x i32> <i32 0, i32 3, i32 2, i32 1>
  ret <4 x double> %B
}