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
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
|
//====- X86InstrFMA.td - Describe the X86 Instruction Set --*- tablegen -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file describes FMA (Fused Multiply-Add) instructions.
//
//===----------------------------------------------------------------------===//
//===----------------------------------------------------------------------===//
// FMA3 - Intel 3 operand Fused Multiply-Add instructions
//===----------------------------------------------------------------------===//
multiclass fma_rm<bits<8> opc, string OpcodeStr> {
def r : FMA3<opc, MRMSrcReg, (outs VR128:$dst),
(ins VR128:$src1, VR128:$src2),
!strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"),
[]>;
def m : FMA3<opc, MRMSrcMem, (outs VR128:$dst),
(ins VR128:$src1, f128mem:$src2),
!strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"),
[]>;
def rY : FMA3<opc, MRMSrcReg, (outs VR256:$dst),
(ins VR256:$src1, VR256:$src2),
!strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"),
[]>;
def mY : FMA3<opc, MRMSrcMem, (outs VR256:$dst),
(ins VR256:$src1, f256mem:$src2),
!strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"),
[]>;
}
multiclass fma_forms<bits<8> opc132, bits<8> opc213, bits<8> opc231,
string OpcodeStr, string PackTy> {
defm r132 : fma_rm<opc132, !strconcat(OpcodeStr, !strconcat("132", PackTy))>;
defm r213 : fma_rm<opc213, !strconcat(OpcodeStr, !strconcat("213", PackTy))>;
defm r231 : fma_rm<opc231, !strconcat(OpcodeStr, !strconcat("231", PackTy))>;
}
let isAsmParserOnly = 1 in {
// Fused Multiply-Add
defm VFMADDPS : fma_forms<0x98, 0xA8, 0xB8, "vfmadd", "ps">;
defm VFMADDPD : fma_forms<0x98, 0xA8, 0xB8, "vfmadd", "pd">, VEX_W;
defm VFMADDSUBPS : fma_forms<0x96, 0xA6, 0xB6, "vfmaddsub", "ps">;
defm VFMADDSUBPD : fma_forms<0x96, 0xA6, 0xB6, "vfmaddsub", "pd">, VEX_W;
defm VFMSUBADDPS : fma_forms<0x97, 0xA7, 0xB7, "vfmsubadd", "ps">;
defm VFMSUBADDPD : fma_forms<0x97, 0xA7, 0xB7, "vfmsubadd", "pd">, VEX_W;
defm VFMSUBPS : fma_forms<0x9A, 0xAA, 0xBA, "vfmsub", "ps">;
defm VFMSUBPD : fma_forms<0x9A, 0xAA, 0xBA, "vfmsub", "pd">, VEX_W;
// Fused Negative Multiply-Add
defm VFNMADDPS : fma_forms<0x9C, 0xAC, 0xBC, "vfnmadd", "ps">;
defm VFNMADDPD : fma_forms<0x9C, 0xAC, 0xBC, "vfnmadd", "pd">, VEX_W;
defm VFNMSUBPS : fma_forms<0x9E, 0xAE, 0xBE, "vfnmsub", "ps">;
defm VFNMSUBPD : fma_forms<0x9E, 0xAE, 0xBE, "vfnmsub", "pd">, VEX_W;
}
//===----------------------------------------------------------------------===//
// FMA4 - AMD 4 operand Fused Multiply-Add instructions
//===----------------------------------------------------------------------===//
multiclass fma4s<bits<8> opc, string OpcodeStr> {
def rr : FMA4<opc, MRMSrcReg, (outs VR128:$dst),
(ins VR128:$src1, VR128:$src2, VR128:$src3),
!strconcat(OpcodeStr,
"\t{$src2, $src3, $src1, $dst|$dst, $src1, $src3, $src2}"),
[]>, XOP_W;
def rm : FMA4<opc, MRMSrcMem, (outs VR128:$dst),
(ins VR128:$src1, VR128:$src2, f128mem:$src3),
!strconcat(OpcodeStr,
"\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}"),
[]>, XOP_W;
def mr : FMA4<opc, MRMSrcMem, (outs VR128:$dst),
(ins VR128:$src1, f128mem:$src2, VR128:$src3),
!strconcat(OpcodeStr,
"\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}"),
[]>;
}
multiclass fma4p<bits<8> opc, string OpcodeStr> {
def rr : FMA4<opc, MRMSrcReg, (outs VR128:$dst),
(ins VR128:$src1, VR128:$src2, VR128:$src3),
!strconcat(OpcodeStr,
"\t{$src2, $src3, $src1, $dst|$dst, $src1, $src3, $src2}"),
[]>, XOP_W;
def rm : FMA4<opc, MRMSrcMem, (outs VR128:$dst),
(ins VR128:$src1, VR128:$src2, f128mem:$src3),
!strconcat(OpcodeStr,
"\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}"),
[]>, XOP_W;
def mr : FMA4<opc, MRMSrcMem, (outs VR128:$dst),
(ins VR128:$src1, f128mem:$src2, VR128:$src3),
!strconcat(OpcodeStr,
"\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}"),
[]>;
def rrY : FMA4<opc, MRMSrcReg, (outs VR256:$dst),
(ins VR256:$src1, VR256:$src2, VR256:$src3),
!strconcat(OpcodeStr,
"\t{$src2, $src3, $src1, $dst|$dst, $src1, $src3, $src2}"),
[]>, XOP_W;
def rmY : FMA4<opc, MRMSrcMem, (outs VR256:$dst),
(ins VR256:$src1, VR256:$src2, f256mem:$src3),
!strconcat(OpcodeStr,
"\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}"),
[]>, XOP_W;
def mrY : FMA4<opc, MRMSrcMem, (outs VR256:$dst),
(ins VR256:$src1, f256mem:$src2, VR256:$src3),
!strconcat(OpcodeStr,
"\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}"),
[]>;
}
let isAsmParserOnly = 1 in {
defm VFMADDSS4 : fma4s<0x6A, "vfmaddss">;
defm VFMADDSD4 : fma4s<0x6B, "vfmaddsd">;
defm VFMADDPS4 : fma4p<0x68, "vfmaddps">;
defm VFMADDPD4 : fma4p<0x69, "vfmaddpd">;
defm VFMSUBSS4 : fma4s<0x6E, "vfmsubss">;
defm VFMSUBSD4 : fma4s<0x6F, "vfmsubsd">;
defm VFMSUBPS4 : fma4p<0x6C, "vfmsubps">;
defm VFMSUBPD4 : fma4p<0x6D, "vfmsubpd">;
defm VFNMADDSS4 : fma4s<0x7A, "vfnmaddss">;
defm VFNMADDSD4 : fma4s<0x7B, "vfnmaddsd">;
defm VFNMADDPS4 : fma4p<0x78, "vfnmaddps">;
defm VFNMADDPD4 : fma4p<0x79, "vfnmaddpd">;
defm VFNMSUBSS4 : fma4s<0x7E, "vfnmsubss">;
defm VFNMSUBSD4 : fma4s<0x7F, "vfnmsubsd">;
defm VFNMSUBPS4 : fma4p<0x7C, "vfnmsubps">;
defm VFNMSUBPD4 : fma4p<0x7D, "vfnmsubpd">;
defm VFMADDSUBPS4 : fma4p<0x5C, "vfmaddsubps">;
defm VFMADDSUBPD4 : fma4p<0x5D, "vfmaddsubpd">;
defm VFMSUBADDPS4 : fma4p<0x5E, "vfmsubaddps">;
defm VFMSUBADDPD4 : fma4p<0x5F, "vfmsubaddpd">;
}
// FMA4 Intrinsics patterns
// VFMADD
def : Pat<(int_x86_fma4_vfmadd_ss VR128:$src1, VR128:$src2, VR128:$src3),
(VFMADDSS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_ss VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFMADDSS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_ss VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFMADDSS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_sd VR128:$src1, VR128:$src2, VR128:$src3),
(VFMADDSD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_sd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFMADDSD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_sd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFMADDSD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_ps VR128:$src1, VR128:$src2, VR128:$src3),
(VFMADDPS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_ps VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFMADDPS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_ps VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFMADDPS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_pd VR128:$src1, VR128:$src2, VR128:$src3),
(VFMADDPD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_pd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFMADDPD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_pd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFMADDPD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_ps_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFMADDPS4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_ps_256 VR256:$src1, VR256:$src2,
(alignedloadv8f32 addr:$src3)),
(VFMADDPS4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_ps_256 VR256:$src1,
(alignedloadv8f32 addr:$src2),
VR256:$src3),
(VFMADDPS4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_pd_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFMADDPD4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_pd_256 VR256:$src1, VR256:$src2,
(alignedloadv4f64 addr:$src3)),
(VFMADDPD4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmadd_pd_256 VR256:$src1,
(alignedloadv4f64 addr:$src2),
VR256:$src3),
(VFMADDPD4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
// VFMSUB
def : Pat<(int_x86_fma4_vfmsub_ss VR128:$src1, VR128:$src2, VR128:$src3),
(VFMSUBSS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_ss VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFMSUBSS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_ss VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFMSUBSS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_sd VR128:$src1, VR128:$src2, VR128:$src3),
(VFMSUBSD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_sd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFMSUBSD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_sd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFMSUBSD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_ps VR128:$src1, VR128:$src2, VR128:$src3),
(VFMSUBPS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_ps VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFMSUBPS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_ps VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFMSUBPS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_pd VR128:$src1, VR128:$src2, VR128:$src3),
(VFMSUBPD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_pd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFMSUBPD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_pd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFMSUBPD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_ps_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFMSUBPS4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_ps_256 VR256:$src1, VR256:$src2,
(alignedloadv8f32 addr:$src3)),
(VFMSUBPS4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_ps_256 VR256:$src1,
(alignedloadv8f32 addr:$src2),
VR256:$src3),
(VFMSUBPS4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_pd_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFMSUBPD4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_pd_256 VR256:$src1, VR256:$src2,
(alignedloadv4f64 addr:$src3)),
(VFMSUBPD4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsub_pd_256 VR256:$src1,
(alignedloadv4f64 addr:$src2),
VR256:$src3),
(VFMSUBPD4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
// VFNMADD
def : Pat<(int_x86_fma4_vfnmadd_ss VR128:$src1, VR128:$src2, VR128:$src3),
(VFNMADDSS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_ss VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFNMADDSS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_ss VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFNMADDSS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_sd VR128:$src1, VR128:$src2, VR128:$src3),
(VFNMADDSD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_sd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFNMADDSD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_sd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFNMADDSD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_ps VR128:$src1, VR128:$src2, VR128:$src3),
(VFNMADDPS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_ps VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFNMADDPS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_ps VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFNMADDPS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_pd VR128:$src1, VR128:$src2, VR128:$src3),
(VFNMADDPD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_pd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFNMADDPD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_pd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFNMADDPD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_ps_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFNMADDPS4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_ps_256 VR256:$src1, VR256:$src2,
(alignedloadv8f32 addr:$src3)),
(VFNMADDPS4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_ps_256 VR256:$src1,
(alignedloadv8f32 addr:$src2),
VR256:$src3),
(VFNMADDPS4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_pd_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFNMADDPD4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_pd_256 VR256:$src1, VR256:$src2,
(alignedloadv4f64 addr:$src3)),
(VFNMADDPD4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmadd_pd_256 VR256:$src1,
(alignedloadv4f64 addr:$src2),
VR256:$src3),
(VFNMADDPD4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
// VFNMSUB
def : Pat<(int_x86_fma4_vfnmsub_ss VR128:$src1, VR128:$src2, VR128:$src3),
(VFNMSUBSS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_ss VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFNMSUBSS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_ss VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFNMSUBSS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_sd VR128:$src1, VR128:$src2, VR128:$src3),
(VFNMSUBSD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_sd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFNMSUBSD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_sd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFNMSUBSD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_ps VR128:$src1, VR128:$src2, VR128:$src3),
(VFNMSUBPS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_ps VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFNMSUBPS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_ps VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFNMSUBPS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_pd VR128:$src1, VR128:$src2, VR128:$src3),
(VFNMSUBPD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_pd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFNMSUBPD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_pd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFNMSUBPD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_ps_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFNMSUBPS4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_ps_256 VR256:$src1, VR256:$src2,
(alignedloadv8f32 addr:$src3)),
(VFNMSUBPS4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_ps_256 VR256:$src1,
(alignedloadv8f32 addr:$src2),
VR256:$src3),
(VFNMSUBPS4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_pd_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFNMSUBPD4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_pd_256 VR256:$src1, VR256:$src2,
(alignedloadv4f64 addr:$src3)),
(VFNMSUBPD4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfnmsub_pd_256 VR256:$src1,
(alignedloadv4f64 addr:$src2),
VR256:$src3),
(VFNMSUBPD4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
// VFMADDSUB
def : Pat<(int_x86_fma4_vfmaddsub_ps VR128:$src1, VR128:$src2, VR128:$src3),
(VFMADDSUBPS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_ps VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFMADDSUBPS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_ps VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFMADDSUBPS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_pd VR128:$src1, VR128:$src2, VR128:$src3),
(VFMADDSUBPD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_pd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFMADDSUBPD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_pd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFMADDSUBPD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_ps_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFMADDSUBPS4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_ps_256 VR256:$src1, VR256:$src2,
(alignedloadv8f32 addr:$src3)),
(VFMADDSUBPS4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_ps_256 VR256:$src1,
(alignedloadv8f32 addr:$src2),
VR256:$src3),
(VFMADDSUBPS4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_pd_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFMADDSUBPD4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_pd_256 VR256:$src1, VR256:$src2,
(alignedloadv4f64 addr:$src3)),
(VFMADDSUBPD4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmaddsub_pd_256 VR256:$src1,
(alignedloadv4f64 addr:$src2),
VR256:$src3),
(VFMADDSUBPD4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
// VFMSUBADD
def : Pat<(int_x86_fma4_vfmsubadd_ps VR128:$src1, VR128:$src2, VR128:$src3),
(VFMSUBADDPS4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_ps VR128:$src1, VR128:$src2,
(alignedloadv4f32 addr:$src3)),
(VFMSUBADDPS4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_ps VR128:$src1, (alignedloadv4f32 addr:$src2),
VR128:$src3),
(VFMSUBADDPS4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_pd VR128:$src1, VR128:$src2, VR128:$src3),
(VFMSUBADDPD4rr VR128:$src1, VR128:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_pd VR128:$src1, VR128:$src2,
(alignedloadv2f64 addr:$src3)),
(VFMSUBADDPD4rm VR128:$src1, VR128:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_pd VR128:$src1, (alignedloadv2f64 addr:$src2),
VR128:$src3),
(VFMSUBADDPD4mr VR128:$src1, addr:$src2, VR128:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_ps_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFMSUBADDPS4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_ps_256 VR256:$src1, VR256:$src2,
(alignedloadv8f32 addr:$src3)),
(VFMSUBADDPS4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_ps_256 VR256:$src1,
(alignedloadv8f32 addr:$src2),
VR256:$src3),
(VFMSUBADDPS4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_pd_256 VR256:$src1, VR256:$src2, VR256:$src3),
(VFMSUBADDPD4rrY VR256:$src1, VR256:$src2, VR256:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_pd_256 VR256:$src1, VR256:$src2,
(alignedloadv4f64 addr:$src3)),
(VFMSUBADDPD4rmY VR256:$src1, VR256:$src2, addr:$src3)>;
def : Pat<(int_x86_fma4_vfmsubadd_pd_256 VR256:$src1,
(alignedloadv4f64 addr:$src2),
VR256:$src3),
(VFMSUBADDPD4mrY VR256:$src1, addr:$src2, VR256:$src3)>;
|
