1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512cd,+prefer-256-bit | FileCheck %s --check-prefix=AVX256
3 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512cd,-prefer-256-bit | FileCheck %s --check-prefixes=AVX512,AVX512VL
4 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512cd,+prefer-256-bit | FileCheck %s --check-prefixes=AVX512,AVX512F
5 ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512cd,-prefer-256-bit | FileCheck %s --check-prefixes=AVX512,AVX512F
7 define <8 x i16> @testv8i16(<8 x i16> %in) {
8 ; AVX256-LABEL: testv8i16:
10 ; AVX256-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
11 ; AVX256-NEXT: vplzcntd %ymm0, %ymm0
12 ; AVX256-NEXT: vpmovdw %ymm0, %xmm0
13 ; AVX256-NEXT: vpsubw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
14 ; AVX256-NEXT: vzeroupper
17 ; AVX512VL-LABEL: testv8i16:
19 ; AVX512VL-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
20 ; AVX512VL-NEXT: vplzcntd %ymm0, %ymm0
21 ; AVX512VL-NEXT: vpmovdw %ymm0, %xmm0
22 ; AVX512VL-NEXT: vpsubw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
23 ; AVX512VL-NEXT: vzeroupper
26 ; AVX512F-LABEL: testv8i16:
28 ; AVX512F-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
29 ; AVX512F-NEXT: vplzcntd %zmm0, %zmm0
30 ; AVX512F-NEXT: vpmovdw %zmm0, %ymm0
31 ; AVX512F-NEXT: vpsubw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
32 ; AVX512F-NEXT: vzeroupper
34 %out = call <8 x i16> @llvm.ctlz.v8i16(<8 x i16> %in, i1 false)
38 define <16 x i8> @testv16i8(<16 x i8> %in) {
39 ; AVX256-LABEL: testv16i8:
41 ; AVX256-NEXT: vmovdqa {{.*#+}} xmm1 = [4,3,2,2,1,1,1,1,0,0,0,0,0,0,0,0]
42 ; AVX256-NEXT: vpshufb %xmm0, %xmm1, %xmm2
43 ; AVX256-NEXT: vpsrlw $4, %xmm0, %xmm0
44 ; AVX256-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
45 ; AVX256-NEXT: vpxor %xmm3, %xmm3, %xmm3
46 ; AVX256-NEXT: vpcmpeqb %xmm3, %xmm0, %xmm3
47 ; AVX256-NEXT: vpand %xmm3, %xmm2, %xmm2
48 ; AVX256-NEXT: vpshufb %xmm0, %xmm1, %xmm0
49 ; AVX256-NEXT: vpaddb %xmm0, %xmm2, %xmm0
52 ; AVX512-LABEL: testv16i8:
54 ; AVX512-NEXT: vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
55 ; AVX512-NEXT: vplzcntd %zmm0, %zmm0
56 ; AVX512-NEXT: vpmovdb %zmm0, %xmm0
57 ; AVX512-NEXT: vpsubb {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
58 ; AVX512-NEXT: vzeroupper
60 %out = call <16 x i8> @llvm.ctlz.v16i8(<16 x i8> %in, i1 false)
64 define <16 x i16> @testv16i16(<16 x i16> %in) {
65 ; AVX256-LABEL: testv16i16:
67 ; AVX256-NEXT: vpmovzxwd {{.*#+}} ymm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
68 ; AVX256-NEXT: vplzcntd %ymm1, %ymm1
69 ; AVX256-NEXT: vpmovdw %ymm1, %xmm1
70 ; AVX256-NEXT: vextracti128 $1, %ymm0, %xmm0
71 ; AVX256-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
72 ; AVX256-NEXT: vplzcntd %ymm0, %ymm0
73 ; AVX256-NEXT: vpmovdw %ymm0, %xmm0
74 ; AVX256-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
75 ; AVX256-NEXT: vpsubw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
78 ; AVX512-LABEL: testv16i16:
80 ; AVX512-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
81 ; AVX512-NEXT: vplzcntd %zmm0, %zmm0
82 ; AVX512-NEXT: vpmovdw %zmm0, %ymm0
83 ; AVX512-NEXT: vpsubw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
85 %out = call <16 x i16> @llvm.ctlz.v16i16(<16 x i16> %in, i1 false)
89 define <32 x i8> @testv32i8(<32 x i8> %in) {
90 ; AVX256-LABEL: testv32i8:
92 ; AVX256-NEXT: vbroadcasti128 {{.*#+}} ymm1 = [4,3,2,2,1,1,1,1,0,0,0,0,0,0,0,0,4,3,2,2,1,1,1,1,0,0,0,0,0,0,0,0]
93 ; AVX256-NEXT: # ymm1 = mem[0,1,0,1]
94 ; AVX256-NEXT: vpshufb %ymm0, %ymm1, %ymm2
95 ; AVX256-NEXT: vpsrlw $4, %ymm0, %ymm0
96 ; AVX256-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to8}, %ymm0, %ymm0
97 ; AVX256-NEXT: vpxor %xmm3, %xmm3, %xmm3
98 ; AVX256-NEXT: vpcmpeqb %ymm3, %ymm0, %ymm3
99 ; AVX256-NEXT: vpand %ymm3, %ymm2, %ymm2
100 ; AVX256-NEXT: vpshufb %ymm0, %ymm1, %ymm0
101 ; AVX256-NEXT: vpaddb %ymm0, %ymm2, %ymm0
104 ; AVX512-LABEL: testv32i8:
106 ; AVX512-NEXT: vpmovzxbd {{.*#+}} zmm1 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
107 ; AVX512-NEXT: vplzcntd %zmm1, %zmm1
108 ; AVX512-NEXT: vpmovdb %zmm1, %xmm1
109 ; AVX512-NEXT: vextracti128 $1, %ymm0, %xmm0
110 ; AVX512-NEXT: vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero
111 ; AVX512-NEXT: vplzcntd %zmm0, %zmm0
112 ; AVX512-NEXT: vpmovdb %zmm0, %xmm0
113 ; AVX512-NEXT: vinserti128 $1, %xmm0, %ymm1, %ymm0
114 ; AVX512-NEXT: vpsubb {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
116 %out = call <32 x i8> @llvm.ctlz.v32i8(<32 x i8> %in, i1 false)
120 declare <8 x i16> @llvm.ctlz.v8i16(<8 x i16>, i1)
121 declare <16 x i8> @llvm.ctlz.v16i8(<16 x i8>, i1)
122 declare <16 x i16> @llvm.ctlz.v16i16(<16 x i16>, i1)
123 declare <32 x i8> @llvm.ctlz.v32i8(<32 x i8>, i1)