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
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 6
; RUN: llc < %s -mtriple=x86_64-- -mattr=+sse2 -verify-machineinstrs | FileCheck %s --check-prefixes=SSE
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx -verify-machineinstrs | FileCheck %s --check-prefixes=AVX,AVX1
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx2 -verify-machineinstrs | FileCheck %s --check-prefixes=AVX,AVX2
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512f,+avx512vl -verify-machineinstrs | FileCheck %s --check-prefixes=AVX512,AVX512VL
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512f,+avx512vl,+avx512bw -verify-machineinstrs | FileCheck %s --check-prefixes=AVX512,AVX512BW
define void @zero_xmm(<4 x i32> %arg) #0 {
; SSE-LABEL: zero_xmm:
; SSE: # %bb.0:
; SSE-NEXT: movaps %xmm0, 0
; SSE-NEXT: xorps %xmm0, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: zero_xmm:
; AVX: # %bb.0:
; AVX-NEXT: vmovaps %xmm0, 0
; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX-NEXT: retq
;
; AVX512-LABEL: zero_xmm:
; AVX512: # %bb.0:
; AVX512-NEXT: vmovaps %xmm0, 0
; AVX512-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX512-NEXT: retq
store <4 x i32> %arg, ptr null, align 32
ret void
}
define void @zero_ymm(<8 x i32> %arg) #0 {
; SSE-LABEL: zero_ymm:
; SSE: # %bb.0:
; SSE-NEXT: movaps %xmm1, 16
; SSE-NEXT: movaps %xmm0, 0
; SSE-NEXT: xorps %xmm0, %xmm0
; SSE-NEXT: xorps %xmm1, %xmm1
; SSE-NEXT: retq
;
; AVX-LABEL: zero_ymm:
; AVX: # %bb.0:
; AVX-NEXT: vmovaps %ymm0, 0
; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX-NEXT: vzeroupper
; AVX-NEXT: retq
;
; AVX512-LABEL: zero_ymm:
; AVX512: # %bb.0:
; AVX512-NEXT: vmovaps %ymm0, 0
; AVX512-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
store <8 x i32> %arg, ptr null, align 32
ret void
}
define void @zero_zmm(<16 x i32> %arg) #0 {
; SSE-LABEL: zero_zmm:
; SSE: # %bb.0:
; SSE-NEXT: movaps %xmm3, 48
; SSE-NEXT: movaps %xmm2, 32
; SSE-NEXT: movaps %xmm1, 16
; SSE-NEXT: movaps %xmm0, 0
; SSE-NEXT: xorps %xmm0, %xmm0
; SSE-NEXT: xorps %xmm1, %xmm1
; SSE-NEXT: xorps %xmm2, %xmm2
; SSE-NEXT: xorps %xmm3, %xmm3
; SSE-NEXT: retq
;
; AVX-LABEL: zero_zmm:
; AVX: # %bb.0:
; AVX-NEXT: vmovaps %ymm1, 32
; AVX-NEXT: vmovaps %ymm0, 0
; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX-NEXT: vxorps %xmm1, %xmm1, %xmm1
; AVX-NEXT: vzeroupper
; AVX-NEXT: retq
;
; AVX512-LABEL: zero_zmm:
; AVX512: # %bb.0:
; AVX512-NEXT: vmovups %zmm0, 0
; AVX512-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX512-NEXT: vzeroupper
; AVX512-NEXT: retq
store <16 x i32> %arg, ptr null, align 32
ret void
}
define void @zero_k(<8 x i32> %arg, <8 x i1> %mask) #0 {
; SSE-LABEL: zero_k:
; SSE: # %bb.0:
; SSE-NEXT: psllw $15, %xmm2
; SSE-NEXT: packsswb %xmm2, %xmm2
; SSE-NEXT: pmovmskb %xmm2, %eax
; SSE-NEXT: testb $1, %al
; SSE-NEXT: jne .LBB3_1
; SSE-NEXT: # %bb.2: # %else
; SSE-NEXT: testb $2, %al
; SSE-NEXT: jne .LBB3_3
; SSE-NEXT: .LBB3_4: # %else2
; SSE-NEXT: testb $4, %al
; SSE-NEXT: jne .LBB3_5
; SSE-NEXT: .LBB3_6: # %else4
; SSE-NEXT: testb $8, %al
; SSE-NEXT: jne .LBB3_7
; SSE-NEXT: .LBB3_8: # %else6
; SSE-NEXT: testb $16, %al
; SSE-NEXT: jne .LBB3_9
; SSE-NEXT: .LBB3_10: # %else8
; SSE-NEXT: testb $32, %al
; SSE-NEXT: jne .LBB3_11
; SSE-NEXT: .LBB3_12: # %else10
; SSE-NEXT: testb $64, %al
; SSE-NEXT: jne .LBB3_13
; SSE-NEXT: .LBB3_14: # %else12
; SSE-NEXT: testb $-128, %al
; SSE-NEXT: je .LBB3_16
; SSE-NEXT: .LBB3_15: # %cond.store13
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm1[3,3,3,3]
; SSE-NEXT: movd %xmm0, 28
; SSE-NEXT: .LBB3_16: # %else14
; SSE-NEXT: xorl %eax, %eax
; SSE-NEXT: pxor %xmm0, %xmm0
; SSE-NEXT: pxor %xmm1, %xmm1
; SSE-NEXT: pxor %xmm2, %xmm2
; SSE-NEXT: retq
; SSE-NEXT: .LBB3_1: # %cond.store
; SSE-NEXT: movd %xmm0, 0
; SSE-NEXT: testb $2, %al
; SSE-NEXT: je .LBB3_4
; SSE-NEXT: .LBB3_3: # %cond.store1
; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,1,1]
; SSE-NEXT: movd %xmm2, 4
; SSE-NEXT: testb $4, %al
; SSE-NEXT: je .LBB3_6
; SSE-NEXT: .LBB3_5: # %cond.store3
; SSE-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
; SSE-NEXT: movd %xmm2, 8
; SSE-NEXT: testb $8, %al
; SSE-NEXT: je .LBB3_8
; SSE-NEXT: .LBB3_7: # %cond.store5
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm0[3,3,3,3]
; SSE-NEXT: movd %xmm0, 12
; SSE-NEXT: testb $16, %al
; SSE-NEXT: je .LBB3_10
; SSE-NEXT: .LBB3_9: # %cond.store7
; SSE-NEXT: movd %xmm1, 16
; SSE-NEXT: testb $32, %al
; SSE-NEXT: je .LBB3_12
; SSE-NEXT: .LBB3_11: # %cond.store9
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; SSE-NEXT: movd %xmm0, 20
; SSE-NEXT: testb $64, %al
; SSE-NEXT: je .LBB3_14
; SSE-NEXT: .LBB3_13: # %cond.store11
; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
; SSE-NEXT: movd %xmm0, 24
; SSE-NEXT: testb $-128, %al
; SSE-NEXT: jne .LBB3_15
; SSE-NEXT: jmp .LBB3_16
;
; AVX1-LABEL: zero_k:
; AVX1: # %bb.0:
; AVX1-NEXT: vpmovzxwd {{.*#+}} xmm2 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
; AVX1-NEXT: vpslld $31, %xmm2, %xmm2
; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm1[4,4,5,5,6,6,7,7]
; AVX1-NEXT: vpslld $31, %xmm1, %xmm1
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm2, %ymm1
; AVX1-NEXT: vmaskmovps %ymm0, %ymm1, 0
; AVX1-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX1-NEXT: vxorps %xmm1, %xmm1, %xmm1
; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
; AVX1-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX1-NEXT: vxorps %xmm1, %xmm1, %xmm1
; AVX1-NEXT: vpxor %xmm2, %xmm2, %xmm2
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
; AVX2-LABEL: zero_k:
; AVX2: # %bb.0:
; AVX2-NEXT: vpmovzxwd {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
; AVX2-NEXT: vpslld $31, %ymm1, %ymm1
; AVX2-NEXT: vpmaskmovd %ymm0, %ymm1, 0
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX2-NEXT: vpxor %xmm0, %xmm0, %xmm0
; AVX2-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
; AVX512VL-LABEL: zero_k:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vpmovsxwd %xmm1, %ymm1
; AVX512VL-NEXT: vpslld $31, %ymm1, %ymm1
; AVX512VL-NEXT: vptestmd %ymm1, %ymm1, %k1
; AVX512VL-NEXT: vmovdqa32 %ymm0, 0 {%k1}
; AVX512VL-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512VL-NEXT: vpxor %xmm0, %xmm0, %xmm0
; AVX512VL-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512VL-NEXT: kxorw %k0, %k0, %k1
; AVX512VL-NEXT: vzeroupper
; AVX512VL-NEXT: retq
;
; AVX512BW-LABEL: zero_k:
; AVX512BW: # %bb.0:
; AVX512BW-NEXT: vpsllw $15, %xmm1, %xmm1
; AVX512BW-NEXT: vpmovw2m %xmm1, %k1
; AVX512BW-NEXT: vmovdqa32 %ymm0, 0 {%k1}
; AVX512BW-NEXT: vpxor %xmm1, %xmm1, %xmm1
; AVX512BW-NEXT: vpxor %xmm0, %xmm0, %xmm0
; AVX512BW-NEXT: kxorq %k0, %k0, %k1
; AVX512BW-NEXT: vzeroupper
; AVX512BW-NEXT: retq
tail call void @llvm.masked.store.v8i32.p0(<8 x i32> %arg, ptr null, i32 32, <8 x i1> %mask)
ret void
}
attributes #0 = { "zero-call-used-regs"="used" }
|
