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
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -O3 -disable-peephole -mtriple=x86_64-unknown-unknown -mattr=+avx512vp2intersect,+avx512vl < %s | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-unknown"
define void @stack_fold_vp2intersectd(<16 x i32>* %a, <16 x i32> %b, <16 x i1>* nocapture %m0, <16 x i1>* nocapture %m1) {
; CHECK-LABEL: stack_fold_vp2intersectd:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %zmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 64-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps (%rdi), %zmm0
; CHECK-NEXT: vp2intersectd {{[-0-9]+}}(%r{{[sb]}}p), %zmm0, %k0 # 64-byte Folded Reload
; CHECK-NEXT: kmovw %k0, (%rsi)
; CHECK-NEXT: kmovw %k1, (%rdx)
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <16 x i32>, <16 x i32>* %a
%3 = tail call { <16 x i1>, <16 x i1> } @llvm.x86.avx512.vp2intersect.d.512(<16 x i32> %2, <16 x i32> %b)
%4 = extractvalue { <16 x i1>, <16 x i1> } %3, 0
store <16 x i1> %4, <16 x i1>* %m0
%5 = extractvalue { <16 x i1>, <16 x i1> } %3, 1
store <16 x i1> %5, <16 x i1>* %m1
ret void
}
declare { <16 x i1>, <16 x i1> } @llvm.x86.avx512.vp2intersect.d.512(<16 x i32>, <16 x i32>)
define void @stack_fold_vp2intersectq(<8 x i64>* %a, <8 x i64> %b, <8 x i1>* nocapture %m0, <8 x i1>* nocapture %m1) {
; CHECK-LABEL: stack_fold_vp2intersectq:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %zmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 64-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps (%rdi), %zmm0
; CHECK-NEXT: vp2intersectq {{[-0-9]+}}(%r{{[sb]}}p), %zmm0, %k0 # 64-byte Folded Reload
; CHECK-NEXT: kmovw %k1, %eax
; CHECK-NEXT: kmovw %k0, %ecx
; CHECK-NEXT: movb %cl, (%rsi)
; CHECK-NEXT: movb %al, (%rdx)
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <8 x i64>, <8 x i64>* %a
%3 = tail call { <8 x i1>, <8 x i1> } @llvm.x86.avx512.vp2intersect.q.512(<8 x i64> %2, <8 x i64> %b)
%4 = extractvalue { <8 x i1>, <8 x i1> } %3, 0
store <8 x i1> %4, <8 x i1>* %m0
%5 = extractvalue { <8 x i1>, <8 x i1> } %3, 1
store <8 x i1> %5, <8 x i1>* %m1
ret void
}
declare { <8 x i1>, <8 x i1> } @llvm.x86.avx512.vp2intersect.q.512(<8 x i64>, <8 x i64>)
define void @stack_fold_vp2intersectd_256(<8 x i32>* %a, <8 x i32> %b, <8 x i1>* nocapture %m0, <8 x i1>* nocapture %m1) {
; CHECK-LABEL: stack_fold_vp2intersectd_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps (%rdi), %ymm0
; CHECK-NEXT: vp2intersectd {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %k0 # 32-byte Folded Reload
; CHECK-NEXT: kmovw %k1, %eax
; CHECK-NEXT: kmovw %k0, %ecx
; CHECK-NEXT: movb %cl, (%rsi)
; CHECK-NEXT: movb %al, (%rdx)
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <8 x i32>, <8 x i32>* %a
%3 = tail call { <8 x i1>, <8 x i1> } @llvm.x86.avx512.vp2intersect.d.256(<8 x i32> %2, <8 x i32> %b)
%4 = extractvalue { <8 x i1>, <8 x i1> } %3, 0
store <8 x i1> %4, <8 x i1>* %m0
%5 = extractvalue { <8 x i1>, <8 x i1> } %3, 1
store <8 x i1> %5, <8 x i1>* %m1
ret void
}
declare { <8 x i1>, <8 x i1> } @llvm.x86.avx512.vp2intersect.d.256(<8 x i32>, <8 x i32>)
define void @stack_fold_vp2intersectq_256(<4 x i64>* %a, <4 x i64> %b, <4 x i1>* nocapture %m0, <4 x i1>* nocapture %m1) {
; CHECK-LABEL: stack_fold_vp2intersectq_256:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovups %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps (%rdi), %ymm0
; CHECK-NEXT: vp2intersectq {{[-0-9]+}}(%r{{[sb]}}p), %ymm0, %k0 # 32-byte Folded Reload
; CHECK-NEXT: kshiftlw $12, %k0, %k2
; CHECK-NEXT: kshiftrw $12, %k2, %k2
; CHECK-NEXT: kmovw %k2, %eax
; CHECK-NEXT: movb %al, (%rsi)
; CHECK-NEXT: kshiftlw $12, %k1, %k0
; CHECK-NEXT: kshiftrw $12, %k0, %k0
; CHECK-NEXT: kmovw %k0, %eax
; CHECK-NEXT: movb %al, (%rdx)
; CHECK-NEXT: vzeroupper
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <4 x i64>, <4 x i64>* %a
%3 = tail call { <4 x i1>, <4 x i1> } @llvm.x86.avx512.vp2intersect.q.256(<4 x i64> %2, <4 x i64> %b)
%4 = extractvalue { <4 x i1>, <4 x i1> } %3, 0
store <4 x i1> %4, <4 x i1>* %m0
%5 = extractvalue { <4 x i1>, <4 x i1> } %3, 1
store <4 x i1> %5, <4 x i1>* %m1
ret void
}
declare { <4 x i1>, <4 x i1> } @llvm.x86.avx512.vp2intersect.q.256(<4 x i64>, <4 x i64>)
define void @stack_fold_vp2intersectd_128(<4 x i32>* %a, <4 x i32> %b, <4 x i1>* nocapture %m0, <4 x i1>* nocapture %m1) {
; CHECK-LABEL: stack_fold_vp2intersectd_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps (%rdi), %xmm0
; CHECK-NEXT: vp2intersectd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %k0 # 16-byte Folded Reload
; CHECK-NEXT: kshiftlw $12, %k0, %k2
; CHECK-NEXT: kshiftrw $12, %k2, %k2
; CHECK-NEXT: kmovw %k2, %eax
; CHECK-NEXT: movb %al, (%rsi)
; CHECK-NEXT: kshiftlw $12, %k1, %k0
; CHECK-NEXT: kshiftrw $12, %k0, %k0
; CHECK-NEXT: kmovw %k0, %eax
; CHECK-NEXT: movb %al, (%rdx)
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <4 x i32>, <4 x i32>* %a
%3 = tail call { <4 x i1>, <4 x i1> } @llvm.x86.avx512.vp2intersect.d.128(<4 x i32> %2, <4 x i32> %b)
%4 = extractvalue { <4 x i1>, <4 x i1> } %3, 0
store <4 x i1> %4, <4 x i1>* %m0
%5 = extractvalue { <4 x i1>, <4 x i1> } %3, 1
store <4 x i1> %5, <4 x i1>* %m1
ret void
}
declare { <4 x i1>, <4 x i1> } @llvm.x86.avx512.vp2intersect.d.128(<4 x i32>, <4 x i32>)
define void @stack_fold_vp2intersectq_128(<2 x i64>* %a, <2 x i64> %b, <2 x i1>* nocapture %m0, <2 x i1>* nocapture %m1) {
; CHECK-LABEL: stack_fold_vp2intersectq_128:
; CHECK: # %bb.0:
; CHECK-NEXT: vmovaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: #APP
; CHECK-NEXT: nop
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: vmovaps (%rdi), %xmm0
; CHECK-NEXT: vp2intersectq {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %k0 # 16-byte Folded Reload
; CHECK-NEXT: kshiftlw $14, %k0, %k2
; CHECK-NEXT: kshiftrw $14, %k2, %k2
; CHECK-NEXT: kmovw %k2, %eax
; CHECK-NEXT: movb %al, (%rsi)
; CHECK-NEXT: kshiftlw $14, %k1, %k0
; CHECK-NEXT: kshiftrw $14, %k0, %k0
; CHECK-NEXT: kmovw %k0, %eax
; CHECK-NEXT: movb %al, (%rdx)
; CHECK-NEXT: retq
%1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
%2 = load <2 x i64>, <2 x i64>* %a
%3 = tail call { <2 x i1>, <2 x i1> } @llvm.x86.avx512.vp2intersect.q.128(<2 x i64> %2, <2 x i64> %b)
%4 = extractvalue { <2 x i1>, <2 x i1> } %3, 0
store <2 x i1> %4, <2 x i1>* %m0
%5 = extractvalue { <2 x i1>, <2 x i1> } %3, 1
store <2 x i1> %5, <2 x i1>* %m1
ret void
}
declare { <2 x i1>, <2 x i1> } @llvm.x86.avx512.vp2intersect.q.128(<2 x i64>, <2 x i64>)
|
