4
4
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512F
5
5
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vl | FileCheck %s --check-prefixes=AVX512VL
6
6
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512BW
7
- ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2 | FileCheck %s --check-prefixes=AVX512VBMI2
7
+ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi,+ avx512vbmi2 | FileCheck %s --check-prefixes=AVX512VBMI2
8
8
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512bw,+avx512vl | FileCheck %s --check-prefixes=AVX512VLBW
9
- ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512VLVBMI2
9
+ ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vbmi,+ avx512vbmi2,+avx512vl | FileCheck %s --check-prefixes=AVX512VLVBMI2
10
10
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx | FileCheck %s --check-prefixes=XOPAVX1
11
11
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+xop,+avx2 | FileCheck %s --check-prefixes=XOPAVX2
12
12
@@ -701,13 +701,13 @@ define <32 x i8> @var_funnnel_v32i8(<32 x i8> %x, <32 x i8> %y, <32 x i8> %amt)
701
701
;
702
702
; AVX512VBMI2-LABEL: var_funnnel_v32i8:
703
703
; AVX512VBMI2: # %bb.0:
704
- ; AVX512VBMI2-NEXT: vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
704
+ ; AVX512VBMI2-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
705
+ ; AVX512VBMI2-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
706
+ ; AVX512VBMI2-NEXT: vmovdqa64 {{.*#+}} zmm3 = [0,64,1,65,2,66,3,67,4,68,5,69,6,70,7,71,8,72,9,73,10,74,11,75,12,76,13,77,14,78,15,79,16,80,17,81,18,82,19,83,20,84,21,85,22,86,23,87,24,88,25,89,26,90,27,91,28,92,29,93,30,94,31,95]
707
+ ; AVX512VBMI2-NEXT: vpermi2b %zmm0, %zmm1, %zmm3
708
+ ; AVX512VBMI2-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm0
705
709
; AVX512VBMI2-NEXT: vpmovzxbw {{.*#+}} 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,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
706
- ; AVX512VBMI2-NEXT: vpsllw $8, %zmm0, %zmm0
707
- ; AVX512VBMI2-NEXT: vporq %zmm1, %zmm0, %zmm0
708
- ; AVX512VBMI2-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm1
709
- ; AVX512VBMI2-NEXT: vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
710
- ; AVX512VBMI2-NEXT: vpsllvw %zmm1, %zmm0, %zmm0
710
+ ; AVX512VBMI2-NEXT: vpsllvw %zmm0, %zmm3, %zmm0
711
711
; AVX512VBMI2-NEXT: vpsrlw $8, %zmm0, %zmm0
712
712
; AVX512VBMI2-NEXT: vpmovwb %zmm0, %ymm0
713
713
; AVX512VBMI2-NEXT: retq
@@ -727,13 +727,13 @@ define <32 x i8> @var_funnnel_v32i8(<32 x i8> %x, <32 x i8> %y, <32 x i8> %amt)
727
727
;
728
728
; AVX512VLVBMI2-LABEL: var_funnnel_v32i8:
729
729
; AVX512VLVBMI2: # %bb.0:
730
- ; AVX512VLVBMI2-NEXT: vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
730
+ ; AVX512VLVBMI2-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
731
+ ; AVX512VLVBMI2-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
732
+ ; AVX512VLVBMI2-NEXT: vmovdqa64 {{.*#+}} zmm3 = [0,64,1,65,2,66,3,67,4,68,5,69,6,70,7,71,8,72,9,73,10,74,11,75,12,76,13,77,14,78,15,79,16,80,17,81,18,82,19,83,20,84,21,85,22,86,23,87,24,88,25,89,26,90,27,91,28,92,29,93,30,94,31,95]
733
+ ; AVX512VLVBMI2-NEXT: vpermi2b %zmm0, %zmm1, %zmm3
734
+ ; AVX512VLVBMI2-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm0
731
735
; AVX512VLVBMI2-NEXT: vpmovzxbw {{.*#+}} 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,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
732
- ; AVX512VLVBMI2-NEXT: vpsllw $8, %zmm0, %zmm0
733
- ; AVX512VLVBMI2-NEXT: vporq %zmm1, %zmm0, %zmm0
734
- ; AVX512VLVBMI2-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm1
735
- ; AVX512VLVBMI2-NEXT: vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
736
- ; AVX512VLVBMI2-NEXT: vpsllvw %zmm1, %zmm0, %zmm0
736
+ ; AVX512VLVBMI2-NEXT: vpsllvw %zmm0, %zmm3, %zmm0
737
737
; AVX512VLVBMI2-NEXT: vpsrlw $8, %zmm0, %zmm0
738
738
; AVX512VLVBMI2-NEXT: vpmovwb %zmm0, %ymm0
739
739
; AVX512VLVBMI2-NEXT: retq
@@ -1994,11 +1994,11 @@ define <32 x i8> @constant_funnnel_v32i8(<32 x i8> %x, <32 x i8> %y) nounwind {
1994
1994
;
1995
1995
; AVX512VBMI2-LABEL: constant_funnnel_v32i8:
1996
1996
; AVX512VBMI2: # %bb.0:
1997
- ; AVX512VBMI2-NEXT: vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero, ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
1998
- ; AVX512VBMI2-NEXT: vpmovzxbw {{.*#+}} 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,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
1999
- ; AVX512VBMI2-NEXT: vpsllw $8, %zmm0, %zmm0
2000
- ; AVX512VBMI2-NEXT: vporq %zmm1 , %zmm0 , %zmm0
2001
- ; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0 , %zmm0
1997
+ ; AVX512VBMI2-NEXT: # kill: def $ ymm1 killed $ ymm1 def $zmm1
1998
+ ; AVX512VBMI2-NEXT: # kill: def $ ymm0 killed $ ymm0 def $zmm0
1999
+ ; AVX512VBMI2-NEXT: vmovdqa64 {{.*#+}} zmm2 = [0,64,1,65,2,66,3,67,4,68,5,69,6,70,7,71,8,72,9,73,10,74,11,75,12,76,13,77,14,78,15,79,16,80,17,81,18,82,19,83,20,84,21,85,22,86,23,87,24,88,25,89,26,90,27,91,28,92,29,93,30,94,31,95]
2000
+ ; AVX512VBMI2-NEXT: vpermi2b %zmm0 , %zmm1 , %zmm2
2001
+ ; AVX512VBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2 , %zmm0
2002
2002
; AVX512VBMI2-NEXT: vpsrlw $8, %zmm0, %zmm0
2003
2003
; AVX512VBMI2-NEXT: vpmovwb %zmm0, %ymm0
2004
2004
; AVX512VBMI2-NEXT: retq
@@ -2016,11 +2016,11 @@ define <32 x i8> @constant_funnnel_v32i8(<32 x i8> %x, <32 x i8> %y) nounwind {
2016
2016
;
2017
2017
; AVX512VLVBMI2-LABEL: constant_funnnel_v32i8:
2018
2018
; AVX512VLVBMI2: # %bb.0:
2019
- ; AVX512VLVBMI2-NEXT: vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero, ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
2020
- ; AVX512VLVBMI2-NEXT: vpmovzxbw {{.*#+}} 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,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
2021
- ; AVX512VLVBMI2-NEXT: vpsllw $8, %zmm0, %zmm0
2022
- ; AVX512VLVBMI2-NEXT: vporq %zmm1 , %zmm0 , %zmm0
2023
- ; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0 , %zmm0
2019
+ ; AVX512VLVBMI2-NEXT: # kill: def $ ymm1 killed $ ymm1 def $zmm1
2020
+ ; AVX512VLVBMI2-NEXT: # kill: def $ ymm0 killed $ ymm0 def $zmm0
2021
+ ; AVX512VLVBMI2-NEXT: vmovdqa64 {{.*#+}} zmm2 = [0,64,1,65,2,66,3,67,4,68,5,69,6,70,7,71,8,72,9,73,10,74,11,75,12,76,13,77,14,78,15,79,16,80,17,81,18,82,19,83,20,84,21,85,22,86,23,87,24,88,25,89,26,90,27,91,28,92,29,93,30,94,31,95]
2022
+ ; AVX512VLVBMI2-NEXT: vpermi2b %zmm0 , %zmm1 , %zmm2
2023
+ ; AVX512VLVBMI2-NEXT: vpsllvw {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2 , %zmm0
2024
2024
; AVX512VLVBMI2-NEXT: vpsrlw $8, %zmm0, %zmm0
2025
2025
; AVX512VLVBMI2-NEXT: vpmovwb %zmm0, %ymm0
2026
2026
; AVX512VLVBMI2-NEXT: retq
0 commit comments