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
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown -mattr=+sse4.1 -show-mc-encoding | FileCheck %s --check-prefixes=SSE,X32-SSE
; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx -show-mc-encoding | FileCheck %s --check-prefixes=AVX,X32-AVX
; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=AVX512VL,X32-AVX512VL
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse4.1 -show-mc-encoding | FileCheck %s --check-prefixes=SSE,X64-SSE
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx -show-mc-encoding | FileCheck %s --check-prefixes=AVX,X64-AVX
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=AVX512VL,X64-AVX512VL
define <2 x double> @fpext_4f32_to_2f64(<4 x float> %a) {
; SSE-LABEL: fpext_4f32_to_2f64:
; SSE: # %bb.0:
; SSE-NEXT: cvtps2pd %xmm0, %xmm0 # encoding: [0x0f,0x5a,0xc0]
; SSE-NEXT: ret{{[l|q]}} # encoding: [0xc3]
;
; AVX-LABEL: fpext_4f32_to_2f64:
; AVX: # %bb.0:
; AVX-NEXT: vcvtps2pd %xmm0, %xmm0 # encoding: [0xc5,0xf8,0x5a,0xc0]
; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3]
;
; AVX512VL-LABEL: fpext_4f32_to_2f64:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vcvtps2pd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x5a,0xc0]
; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%cvt = fpext <4 x float> %a to <4 x double>
%shuf = shufflevector <4 x double> %cvt, <4 x double> undef, <2 x i32> <i32 0, i32 1>
ret <2 x double> %shuf
}
define <2 x double> @fpext_8f32_to_2f64(<8 x float> %a) {
; SSE-LABEL: fpext_8f32_to_2f64:
; SSE: # %bb.0:
; SSE-NEXT: cvtps2pd %xmm0, %xmm0 # encoding: [0x0f,0x5a,0xc0]
; SSE-NEXT: ret{{[l|q]}} # encoding: [0xc3]
;
; AVX-LABEL: fpext_8f32_to_2f64:
; AVX: # %bb.0:
; AVX-NEXT: vcvtps2pd %xmm0, %xmm0 # encoding: [0xc5,0xf8,0x5a,0xc0]
; AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3]
;
; AVX512VL-LABEL: fpext_8f32_to_2f64:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vcvtps2pd %ymm0, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x5a,0xc0]
; AVX512VL-NEXT: # kill: def $xmm0 killed $xmm0 killed $zmm0
; AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%cvt = fpext <8 x float> %a to <8 x double>
%shuf = shufflevector <8 x double> %cvt, <8 x double> undef, <2 x i32> <i32 0, i32 1>
ret <2 x double> %shuf
}
define <4 x double> @fpext_8f32_to_4f64(<8 x float> %a) {
; SSE-LABEL: fpext_8f32_to_4f64:
; SSE: # %bb.0:
; SSE-NEXT: cvtps2pd %xmm0, %xmm2 # encoding: [0x0f,0x5a,0xd0]
; SSE-NEXT: movhlps %xmm0, %xmm0 # encoding: [0x0f,0x12,0xc0]
; SSE-NEXT: # xmm0 = xmm0[1,1]
; SSE-NEXT: cvtps2pd %xmm0, %xmm1 # encoding: [0x0f,0x5a,0xc8]
; SSE-NEXT: movaps %xmm2, %xmm0 # encoding: [0x0f,0x28,0xc2]
; SSE-NEXT: ret{{[l|q]}} # encoding: [0xc3]
;
; AVX-LABEL: fpext_8f32_to_4f64:
; AVX: # %bb.0:
; AVX-NEXT: vcvtps2pd %xmm0, %ymm0 # encoding: [0xc5,0xfc,0x5a,0xc0]
; AVX-NEXT: ret{{[l|q]}} # encoding: [0xc3]
;
; AVX512VL-LABEL: fpext_8f32_to_4f64:
; AVX512VL: # %bb.0:
; AVX512VL-NEXT: vcvtps2pd %ymm0, %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x5a,0xc0]
; AVX512VL-NEXT: # kill: def $ymm0 killed $ymm0 killed $zmm0
; AVX512VL-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%cvt = fpext <8 x float> %a to <8 x double>
%shuf = shufflevector <8 x double> %cvt, <8 x double> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
ret <4 x double> %shuf
}
; PR11674
define void @fpext_frommem(<2 x float>* %in, <2 x double>* %out) {
; X32-SSE-LABEL: fpext_frommem:
; X32-SSE: # %bb.0: # %entry
; X32-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-SSE-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-SSE-NEXT: cvtps2pd (%ecx), %xmm0 # encoding: [0x0f,0x5a,0x01]
; X32-SSE-NEXT: movups %xmm0, (%eax) # encoding: [0x0f,0x11,0x00]
; X32-SSE-NEXT: retl # encoding: [0xc3]
;
; X32-AVX-LABEL: fpext_frommem:
; X32-AVX: # %bb.0: # %entry
; X32-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-AVX-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-AVX-NEXT: vcvtps2pd (%ecx), %xmm0 # encoding: [0xc5,0xf8,0x5a,0x01]
; X32-AVX-NEXT: vmovups %xmm0, (%eax) # encoding: [0xc5,0xf8,0x11,0x00]
; X32-AVX-NEXT: retl # encoding: [0xc3]
;
; X32-AVX512VL-LABEL: fpext_frommem:
; X32-AVX512VL: # %bb.0: # %entry
; X32-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-AVX512VL-NEXT: vcvtps2pd (%ecx), %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x5a,0x01]
; X32-AVX512VL-NEXT: vmovups %xmm0, (%eax) # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x11,0x00]
; X32-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-SSE-LABEL: fpext_frommem:
; X64-SSE: # %bb.0: # %entry
; X64-SSE-NEXT: cvtps2pd (%rdi), %xmm0 # encoding: [0x0f,0x5a,0x07]
; X64-SSE-NEXT: movups %xmm0, (%rsi) # encoding: [0x0f,0x11,0x06]
; X64-SSE-NEXT: retq # encoding: [0xc3]
;
; X64-AVX-LABEL: fpext_frommem:
; X64-AVX: # %bb.0: # %entry
; X64-AVX-NEXT: vcvtps2pd (%rdi), %xmm0 # encoding: [0xc5,0xf8,0x5a,0x07]
; X64-AVX-NEXT: vmovups %xmm0, (%rsi) # encoding: [0xc5,0xf8,0x11,0x06]
; X64-AVX-NEXT: retq # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: fpext_frommem:
; X64-AVX512VL: # %bb.0: # %entry
; X64-AVX512VL-NEXT: vcvtps2pd (%rdi), %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x5a,0x07]
; X64-AVX512VL-NEXT: vmovups %xmm0, (%rsi) # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x11,0x06]
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
entry:
%0 = load <2 x float>, <2 x float>* %in, align 8
%1 = fpext <2 x float> %0 to <2 x double>
store <2 x double> %1, <2 x double>* %out, align 1
ret void
}
define void @fpext_frommem4(<4 x float>* %in, <4 x double>* %out) {
; X32-SSE-LABEL: fpext_frommem4:
; X32-SSE: # %bb.0: # %entry
; X32-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-SSE-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-SSE-NEXT: cvtps2pd (%ecx), %xmm0 # encoding: [0x0f,0x5a,0x01]
; X32-SSE-NEXT: cvtps2pd 8(%ecx), %xmm1 # encoding: [0x0f,0x5a,0x49,0x08]
; X32-SSE-NEXT: movups %xmm1, 16(%eax) # encoding: [0x0f,0x11,0x48,0x10]
; X32-SSE-NEXT: movups %xmm0, (%eax) # encoding: [0x0f,0x11,0x00]
; X32-SSE-NEXT: retl # encoding: [0xc3]
;
; X32-AVX-LABEL: fpext_frommem4:
; X32-AVX: # %bb.0: # %entry
; X32-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-AVX-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-AVX-NEXT: vcvtps2pd (%ecx), %ymm0 # encoding: [0xc5,0xfc,0x5a,0x01]
; X32-AVX-NEXT: vmovups %ymm0, (%eax) # encoding: [0xc5,0xfc,0x11,0x00]
; X32-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X32-AVX-NEXT: retl # encoding: [0xc3]
;
; X32-AVX512VL-LABEL: fpext_frommem4:
; X32-AVX512VL: # %bb.0: # %entry
; X32-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-AVX512VL-NEXT: vcvtps2pd (%ecx), %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x5a,0x01]
; X32-AVX512VL-NEXT: vmovups %ymm0, (%eax) # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x11,0x00]
; X32-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X32-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-SSE-LABEL: fpext_frommem4:
; X64-SSE: # %bb.0: # %entry
; X64-SSE-NEXT: cvtps2pd (%rdi), %xmm0 # encoding: [0x0f,0x5a,0x07]
; X64-SSE-NEXT: cvtps2pd 8(%rdi), %xmm1 # encoding: [0x0f,0x5a,0x4f,0x08]
; X64-SSE-NEXT: movups %xmm1, 16(%rsi) # encoding: [0x0f,0x11,0x4e,0x10]
; X64-SSE-NEXT: movups %xmm0, (%rsi) # encoding: [0x0f,0x11,0x06]
; X64-SSE-NEXT: retq # encoding: [0xc3]
;
; X64-AVX-LABEL: fpext_frommem4:
; X64-AVX: # %bb.0: # %entry
; X64-AVX-NEXT: vcvtps2pd (%rdi), %ymm0 # encoding: [0xc5,0xfc,0x5a,0x07]
; X64-AVX-NEXT: vmovups %ymm0, (%rsi) # encoding: [0xc5,0xfc,0x11,0x06]
; X64-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-AVX-NEXT: retq # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: fpext_frommem4:
; X64-AVX512VL: # %bb.0: # %entry
; X64-AVX512VL-NEXT: vcvtps2pd (%rdi), %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x5a,0x07]
; X64-AVX512VL-NEXT: vmovups %ymm0, (%rsi) # EVEX TO VEX Compression encoding: [0xc5,0xfc,0x11,0x06]
; X64-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
entry:
%0 = load <4 x float>, <4 x float>* %in
%1 = fpext <4 x float> %0 to <4 x double>
store <4 x double> %1, <4 x double>* %out, align 1
ret void
}
define void @fpext_frommem8(<8 x float>* %in, <8 x double>* %out) {
; X32-SSE-LABEL: fpext_frommem8:
; X32-SSE: # %bb.0: # %entry
; X32-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-SSE-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-SSE-NEXT: cvtps2pd 8(%ecx), %xmm0 # encoding: [0x0f,0x5a,0x41,0x08]
; X32-SSE-NEXT: cvtps2pd (%ecx), %xmm1 # encoding: [0x0f,0x5a,0x09]
; X32-SSE-NEXT: cvtps2pd 24(%ecx), %xmm2 # encoding: [0x0f,0x5a,0x51,0x18]
; X32-SSE-NEXT: cvtps2pd 16(%ecx), %xmm3 # encoding: [0x0f,0x5a,0x59,0x10]
; X32-SSE-NEXT: movups %xmm3, 32(%eax) # encoding: [0x0f,0x11,0x58,0x20]
; X32-SSE-NEXT: movups %xmm2, 48(%eax) # encoding: [0x0f,0x11,0x50,0x30]
; X32-SSE-NEXT: movups %xmm1, (%eax) # encoding: [0x0f,0x11,0x08]
; X32-SSE-NEXT: movups %xmm0, 16(%eax) # encoding: [0x0f,0x11,0x40,0x10]
; X32-SSE-NEXT: retl # encoding: [0xc3]
;
; X32-AVX-LABEL: fpext_frommem8:
; X32-AVX: # %bb.0: # %entry
; X32-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-AVX-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-AVX-NEXT: vcvtps2pd (%ecx), %ymm0 # encoding: [0xc5,0xfc,0x5a,0x01]
; X32-AVX-NEXT: vcvtps2pd 16(%ecx), %ymm1 # encoding: [0xc5,0xfc,0x5a,0x49,0x10]
; X32-AVX-NEXT: vmovups %ymm1, 32(%eax) # encoding: [0xc5,0xfc,0x11,0x48,0x20]
; X32-AVX-NEXT: vmovups %ymm0, (%eax) # encoding: [0xc5,0xfc,0x11,0x00]
; X32-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X32-AVX-NEXT: retl # encoding: [0xc3]
;
; X32-AVX512VL-LABEL: fpext_frommem8:
; X32-AVX512VL: # %bb.0: # %entry
; X32-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x08]
; X32-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %ecx # encoding: [0x8b,0x4c,0x24,0x04]
; X32-AVX512VL-NEXT: vcvtps2pd (%ecx), %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x5a,0x01]
; X32-AVX512VL-NEXT: vmovups %zmm0, (%eax) # encoding: [0x62,0xf1,0x7c,0x48,0x11,0x00]
; X32-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X32-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-SSE-LABEL: fpext_frommem8:
; X64-SSE: # %bb.0: # %entry
; X64-SSE-NEXT: cvtps2pd 8(%rdi), %xmm0 # encoding: [0x0f,0x5a,0x47,0x08]
; X64-SSE-NEXT: cvtps2pd (%rdi), %xmm1 # encoding: [0x0f,0x5a,0x0f]
; X64-SSE-NEXT: cvtps2pd 24(%rdi), %xmm2 # encoding: [0x0f,0x5a,0x57,0x18]
; X64-SSE-NEXT: cvtps2pd 16(%rdi), %xmm3 # encoding: [0x0f,0x5a,0x5f,0x10]
; X64-SSE-NEXT: movups %xmm3, 32(%rsi) # encoding: [0x0f,0x11,0x5e,0x20]
; X64-SSE-NEXT: movups %xmm2, 48(%rsi) # encoding: [0x0f,0x11,0x56,0x30]
; X64-SSE-NEXT: movups %xmm1, (%rsi) # encoding: [0x0f,0x11,0x0e]
; X64-SSE-NEXT: movups %xmm0, 16(%rsi) # encoding: [0x0f,0x11,0x46,0x10]
; X64-SSE-NEXT: retq # encoding: [0xc3]
;
; X64-AVX-LABEL: fpext_frommem8:
; X64-AVX: # %bb.0: # %entry
; X64-AVX-NEXT: vcvtps2pd (%rdi), %ymm0 # encoding: [0xc5,0xfc,0x5a,0x07]
; X64-AVX-NEXT: vcvtps2pd 16(%rdi), %ymm1 # encoding: [0xc5,0xfc,0x5a,0x4f,0x10]
; X64-AVX-NEXT: vmovups %ymm1, 32(%rsi) # encoding: [0xc5,0xfc,0x11,0x4e,0x20]
; X64-AVX-NEXT: vmovups %ymm0, (%rsi) # encoding: [0xc5,0xfc,0x11,0x06]
; X64-AVX-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-AVX-NEXT: retq # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: fpext_frommem8:
; X64-AVX512VL: # %bb.0: # %entry
; X64-AVX512VL-NEXT: vcvtps2pd (%rdi), %zmm0 # encoding: [0x62,0xf1,0x7c,0x48,0x5a,0x07]
; X64-AVX512VL-NEXT: vmovups %zmm0, (%rsi) # encoding: [0x62,0xf1,0x7c,0x48,0x11,0x06]
; X64-AVX512VL-NEXT: vzeroupper # encoding: [0xc5,0xf8,0x77]
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
entry:
%0 = load <8 x float>, <8 x float>* %in
%1 = fpext <8 x float> %0 to <8 x double>
store <8 x double> %1, <8 x double>* %out, align 1
ret void
}
define <2 x double> @fpext_fromconst() {
; X32-SSE-LABEL: fpext_fromconst:
; X32-SSE: # %bb.0: # %entry
; X32-SSE-NEXT: movaps {{.*#+}} xmm0 = [1.0E+0,-2.0E+0]
; X32-SSE-NEXT: # encoding: [0x0f,0x28,0x05,A,A,A,A]
; X32-SSE-NEXT: # fixup A - offset: 3, value: {{\.LCPI.*}}, kind: FK_Data_4
; X32-SSE-NEXT: retl # encoding: [0xc3]
;
; X32-AVX-LABEL: fpext_fromconst:
; X32-AVX: # %bb.0: # %entry
; X32-AVX-NEXT: vmovaps {{.*#+}} xmm0 = [1.0E+0,-2.0E+0]
; X32-AVX-NEXT: # encoding: [0xc5,0xf8,0x28,0x05,A,A,A,A]
; X32-AVX-NEXT: # fixup A - offset: 4, value: {{\.LCPI.*}}, kind: FK_Data_4
; X32-AVX-NEXT: retl # encoding: [0xc3]
;
; X32-AVX512VL-LABEL: fpext_fromconst:
; X32-AVX512VL: # %bb.0: # %entry
; X32-AVX512VL-NEXT: vmovaps {{\.LCPI.*}}, %xmm0 # EVEX TO VEX Compression xmm0 = [1.0E+0,-2.0E+0]
; X32-AVX512VL-NEXT: # encoding: [0xc5,0xf8,0x28,0x05,A,A,A,A]
; X32-AVX512VL-NEXT: # fixup A - offset: 4, value: {{\.LCPI.*}}, kind: FK_Data_4
; X32-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-SSE-LABEL: fpext_fromconst:
; X64-SSE: # %bb.0: # %entry
; X64-SSE-NEXT: movaps {{.*#+}} xmm0 = [1.0E+0,-2.0E+0]
; X64-SSE-NEXT: # encoding: [0x0f,0x28,0x05,A,A,A,A]
; X64-SSE-NEXT: # fixup A - offset: 3, value: {{\.LCPI.*}}-4, kind: reloc_riprel_4byte
; X64-SSE-NEXT: retq # encoding: [0xc3]
;
; X64-AVX-LABEL: fpext_fromconst:
; X64-AVX: # %bb.0: # %entry
; X64-AVX-NEXT: vmovaps {{.*#+}} xmm0 = [1.0E+0,-2.0E+0]
; X64-AVX-NEXT: # encoding: [0xc5,0xf8,0x28,0x05,A,A,A,A]
; X64-AVX-NEXT: # fixup A - offset: 4, value: {{\.LCPI.*}}-4, kind: reloc_riprel_4byte
; X64-AVX-NEXT: retq # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: fpext_fromconst:
; X64-AVX512VL: # %bb.0: # %entry
; X64-AVX512VL-NEXT: vmovaps {{.*}}(%rip), %xmm0 # EVEX TO VEX Compression xmm0 = [1.0E+0,-2.0E+0]
; X64-AVX512VL-NEXT: # encoding: [0xc5,0xf8,0x28,0x05,A,A,A,A]
; X64-AVX512VL-NEXT: # fixup A - offset: 4, value: {{\.LCPI.*}}-4, kind: reloc_riprel_4byte
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
entry:
%0 = insertelement <2 x float> undef, float 1.0, i32 0
%1 = insertelement <2 x float> %0, float -2.0, i32 1
%2 = fpext <2 x float> %1 to <2 x double>
ret <2 x double> %2
}
; Make sure we don't narrow a volatile load.
define <2 x double> @PR42079(<4 x float>* %x) {
; X32-SSE-LABEL: PR42079:
; X32-SSE: # %bb.0:
; X32-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X32-SSE-NEXT: movaps (%eax), %xmm0 # encoding: [0x0f,0x28,0x00]
; X32-SSE-NEXT: cvtps2pd %xmm0, %xmm0 # encoding: [0x0f,0x5a,0xc0]
; X32-SSE-NEXT: retl # encoding: [0xc3]
;
; X32-AVX-LABEL: PR42079:
; X32-AVX: # %bb.0:
; X32-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X32-AVX-NEXT: vmovaps (%eax), %xmm0 # encoding: [0xc5,0xf8,0x28,0x00]
; X32-AVX-NEXT: vcvtps2pd %xmm0, %xmm0 # encoding: [0xc5,0xf8,0x5a,0xc0]
; X32-AVX-NEXT: retl # encoding: [0xc3]
;
; X32-AVX512VL-LABEL: PR42079:
; X32-AVX512VL: # %bb.0:
; X32-AVX512VL-NEXT: movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
; X32-AVX512VL-NEXT: vmovaps (%eax), %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0x00]
; X32-AVX512VL-NEXT: vcvtps2pd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x5a,0xc0]
; X32-AVX512VL-NEXT: retl # encoding: [0xc3]
;
; X64-SSE-LABEL: PR42079:
; X64-SSE: # %bb.0:
; X64-SSE-NEXT: movaps (%rdi), %xmm0 # encoding: [0x0f,0x28,0x07]
; X64-SSE-NEXT: cvtps2pd %xmm0, %xmm0 # encoding: [0x0f,0x5a,0xc0]
; X64-SSE-NEXT: retq # encoding: [0xc3]
;
; X64-AVX-LABEL: PR42079:
; X64-AVX: # %bb.0:
; X64-AVX-NEXT: vmovaps (%rdi), %xmm0 # encoding: [0xc5,0xf8,0x28,0x07]
; X64-AVX-NEXT: vcvtps2pd %xmm0, %xmm0 # encoding: [0xc5,0xf8,0x5a,0xc0]
; X64-AVX-NEXT: retq # encoding: [0xc3]
;
; X64-AVX512VL-LABEL: PR42079:
; X64-AVX512VL: # %bb.0:
; X64-AVX512VL-NEXT: vmovaps (%rdi), %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x28,0x07]
; X64-AVX512VL-NEXT: vcvtps2pd %xmm0, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xf8,0x5a,0xc0]
; X64-AVX512VL-NEXT: retq # encoding: [0xc3]
%a = load volatile <4 x float>, <4 x float>* %x
%b = shufflevector <4 x float> %a, <4 x float> %a, <2 x i32> <i32 0, i32 1>
%c = fpext <2 x float> %b to <2 x double>
ret <2 x double> %c
}
|