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
|
;=========================== begin_copyright_notice ============================
;
; Copyright (C) 2025 Intel Corporation
;
; SPDX-License-Identifier: MIT
;
;============================ end_copyright_notice =============================
; REQUIRES: regkeys
; RUN: igc_opt --opaque-pointers --regkey DisableCodeScheduling=0 --regkey EnableCodeSchedulingIfNoSpills=1 \
; RUN: --regkey PrintToConsole=1 --regkey DumpCodeScheduling=1 --igc-code-scheduling -S %s 2>&1 | FileCheck %s
; Checks that the register pressure is estimated correctly for the special cases related to vector shuffles.
define spir_kernel void @vector_shuffle_no_op(ptr addrspace(1) %A) {
; CHECK: Function vector_shuffle_no_op
; CHECK: Greedy MW attempt
; CHECK: {{([0-9]+,[ ]*[0-9]+[ ]*).*[ ]*}} [[BASE_ADDR:%.*]] = ptrtoint ptr addrspace(1) [[A:%.*]] to i64
; (6, 512 ) MW: Node #1, MW: 3000 %load2d = call <16 x i16> @llvm.genx.GenISA.LSC2DBlockRead.v16i16(i64 %base_addr, i32 127, i32 1023, i32 127, i32 0, i32 0, i32 16, i32 16, i32 16, i32 2, i1 false, i1 false, i32 4)
; adds 512 bytes
; CHECK: {{([0-9]+,[ ]*512[ ]*).*[ ]*}} [[LOAD2D:%.*]] = call <16 x i16> @llvm.genx.GenISA.LSC2DBlockRead.v16i16(i64 [[BASE_ADDR]], i32 127, i32 1023, i32 127, i32 0, i32 0, i32 16, i32 16, i32 16, i32 2, i1 false, i1 false, i32 4)
; the EE and IE instructions are marked as NOP and don't add regpressure
; (22, 0 ) Im: NOP Node #2, MW: 3000 %EE1.0 = extractelement <16 x i16> %load2d, i32 0
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE1_0:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 0
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE1_0:%.*]] = insertelement <8 x i16> undef, i16 [[EE1_0]], i32 0
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE1_1:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 1
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE1_1:%.*]] = insertelement <8 x i16> [[IE1_0]], i16 [[EE1_1]], i32 1
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE1_2:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 2
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE1_2:%.*]] = insertelement <8 x i16> [[IE1_1]], i16 [[EE1_2]], i32 2
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE1_3:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 3
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE1_3:%.*]] = insertelement <8 x i16> [[IE1_2]], i16 [[EE1_3]], i32 3
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE1_4:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 4
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE1_4:%.*]] = insertelement <8 x i16> [[IE1_3]], i16 [[EE1_4]], i32 4
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE1_5:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 5
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE1_5:%.*]] = insertelement <8 x i16> [[IE1_4]], i16 [[EE1_5]], i32 5
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE1_6:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE1_6:%.*]] = insertelement <8 x i16> [[IE1_5]], i16 [[EE1_6]], i32 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE1_7:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 7
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE1_7:%.*]] = insertelement <8 x i16> [[IE1_6]], i16 [[EE1_7]], i32 7
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE2_0:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 8
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE2_0:%.*]] = insertelement <8 x i16> undef, i16 [[EE2_0]], i32 0
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE2_1:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 9
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE2_1:%.*]] = insertelement <8 x i16> [[IE2_0]], i16 [[EE2_1]], i32 1
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE2_2:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 10
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE2_2:%.*]] = insertelement <8 x i16> [[IE2_1]], i16 [[EE2_2]], i32 2
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE2_3:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 11
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE2_3:%.*]] = insertelement <8 x i16> [[IE2_2]], i16 [[EE2_3]], i32 3
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE2_4:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 12
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE2_4:%.*]] = insertelement <8 x i16> [[IE2_3]], i16 [[EE2_4]], i32 4
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE2_5:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 13
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE2_5:%.*]] = insertelement <8 x i16> [[IE2_4]], i16 [[EE2_5]], i32 5
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE2_6:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 14
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE2_6:%.*]] = insertelement <8 x i16> [[IE2_5]], i16 [[EE2_6]], i32 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[EE2_7:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 15
; CHECK: {{([0-9]+,[ ]*0[ ]*).*NOP.*[ ]*}} [[IE2_7:%.*]] = insertelement <8 x i16> [[IE2_6]], i16 [[EE2_7]], i32 7
; (22, 512 ) MW: Node #34, MW: 0 %dpas1 = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> %IE1.7, <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
; (38, 0 ) MW: Node #35, MW: 0 %dpas2 = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> %IE2.7, <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
; first DPAS increases the regpressure by 512 bytes, the second one doesn't add any regpressure because the whole vector dies
; CHECK: {{([0-9]+,[ ]*512[ ]*).*[ ]*}} [[DPAS1:%.*]] = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> [[IE1_7]], <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
; CHECK: {{([0-9]+,[ ]*0[ ]*).*[ ]*}} [[DPAS2:%.*]] = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> [[IE2_7]], <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
; CHECK: ret void
;
entry:
%base_addr = ptrtoint ptr addrspace(1) %A to i64
%load2d = call <16 x i16> @llvm.genx.GenISA.LSC2DBlockRead.v16i16(i64 %base_addr, i32 127, i32 1023, i32 127, i32 0, i32 0, i32 16, i32 16, i32 16, i32 2, i1 false, i1 false, i32 4)
%EE1.0 = extractelement <16 x i16> %load2d, i32 0
%IE1.0 = insertelement <8 x i16> undef, i16 %EE1.0, i32 0
%EE1.1 = extractelement <16 x i16> %load2d, i32 1
%IE1.1 = insertelement <8 x i16> %IE1.0, i16 %EE1.1, i32 1
%EE1.2 = extractelement <16 x i16> %load2d, i32 2
%IE1.2 = insertelement <8 x i16> %IE1.1, i16 %EE1.2, i32 2
%EE1.3 = extractelement <16 x i16> %load2d, i32 3
%IE1.3 = insertelement <8 x i16> %IE1.2, i16 %EE1.3, i32 3
%EE1.4 = extractelement <16 x i16> %load2d, i32 4
%IE1.4 = insertelement <8 x i16> %IE1.3, i16 %EE1.4, i32 4
%EE1.5 = extractelement <16 x i16> %load2d, i32 5
%IE1.5 = insertelement <8 x i16> %IE1.4, i16 %EE1.5, i32 5
%EE1.6 = extractelement <16 x i16> %load2d, i32 6
%IE1.6 = insertelement <8 x i16> %IE1.5, i16 %EE1.6, i32 6
%EE1.7 = extractelement <16 x i16> %load2d, i32 7
%IE1.7 = insertelement <8 x i16> %IE1.6, i16 %EE1.7, i32 7
%EE2.0 = extractelement <16 x i16> %load2d, i32 8
%IE2.0 = insertelement <8 x i16> undef, i16 %EE2.0, i32 0
%EE2.1 = extractelement <16 x i16> %load2d, i32 9
%IE2.1 = insertelement <8 x i16> %IE2.0, i16 %EE2.1, i32 1
%EE2.2 = extractelement <16 x i16> %load2d, i32 10
%IE2.2 = insertelement <8 x i16> %IE2.1, i16 %EE2.2, i32 2
%EE2.3 = extractelement <16 x i16> %load2d, i32 11
%IE2.3 = insertelement <8 x i16> %IE2.2, i16 %EE2.3, i32 3
%EE2.4 = extractelement <16 x i16> %load2d, i32 12
%IE2.4 = insertelement <8 x i16> %IE2.3, i16 %EE2.4, i32 4
%EE2.5 = extractelement <16 x i16> %load2d, i32 13
%IE2.5 = insertelement <8 x i16> %IE2.4, i16 %EE2.5, i32 5
%EE2.6 = extractelement <16 x i16> %load2d, i32 14
%IE2.6 = insertelement <8 x i16> %IE2.5, i16 %EE2.6, i32 6
%EE2.7 = extractelement <16 x i16> %load2d, i32 15
%IE2.7 = insertelement <8 x i16> %IE2.6, i16 %EE2.7, i32 7
%dpas1 = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> %IE1.7, <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
%dpas2 = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> %IE2.7, <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
ret void
}
define spir_kernel void @vector_shuffle(ptr addrspace(1) %A) {
; CHECK: Function vector_shuffle
; CHECK: {{([0-9]+,[ ]*[0-9]+[ ]*).*[ ]*}} [[BASE_ADDR:%.*]] = ptrtoint ptr addrspace(1) [[A:%.*]] to i64
; (6, 512 ) MW: Node #1, MW: 3000 %load2d = call <16 x i16> @llvm.genx.GenISA.LSC2DBlockRead.v16i16(i64 %base_addr, i32 127, i32 1023, i32 127, i32 0, i32 0, i32 16, i32 16, i32 16, i32 2, i1 false, i1 false, i32 4)
; adds 512 bytes
; CHECK: {{([0-9]+,[ ]*512[ ]*).*[ ]*}} [[LOAD2D:%.*]] = call <16 x i16> @llvm.genx.GenISA.LSC2DBlockRead.v16i16(i64 [[BASE_ADDR]], i32 127, i32 1023, i32 127, i32 0, i32 0, i32 16, i32 16, i32 16, i32 2, i1 false, i1 false, i32 4)
; the EE and IE instructions are marked as VS. IEs add regpressure. The last IE kills the original vector
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE1_0:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 0
; CHECK: {{([0-9]+,[ ]*256[ ]*).*VS.*[ ]*}} [[IE1_0:%.*]] = insertelement <8 x i16> undef, i16 [[EE1_0]], i32 0
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE1_1:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 2
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE1_1:%.*]] = insertelement <8 x i16> [[IE1_0]], i16 [[EE1_1]], i32 1
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE1_2:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 4
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE1_2:%.*]] = insertelement <8 x i16> [[IE1_1]], i16 [[EE1_2]], i32 2
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE1_3:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE1_3:%.*]] = insertelement <8 x i16> [[IE1_2]], i16 [[EE1_3]], i32 3
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE1_4:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 8
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE1_4:%.*]] = insertelement <8 x i16> [[IE1_3]], i16 [[EE1_4]], i32 4
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE1_5:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 10
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE1_5:%.*]] = insertelement <8 x i16> [[IE1_4]], i16 [[EE1_5]], i32 5
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE1_6:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 12
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE1_6:%.*]] = insertelement <8 x i16> [[IE1_5]], i16 [[EE1_6]], i32 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE1_7:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 14
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE1_7:%.*]] = insertelement <8 x i16> [[IE1_6]], i16 [[EE1_7]], i32 7
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE2_0:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 1
; CHECK: {{([0-9]+,[ ]*256[ ]*).*VS.*[ ]*}} [[IE2_0:%.*]] = insertelement <8 x i16> undef, i16 [[EE2_0]], i32 0
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE2_1:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 3
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE2_1:%.*]] = insertelement <8 x i16> [[IE2_0]], i16 [[EE2_1]], i32 1
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE2_2:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 5
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE2_2:%.*]] = insertelement <8 x i16> [[IE2_1]], i16 [[EE2_2]], i32 2
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE2_3:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 7
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE2_3:%.*]] = insertelement <8 x i16> [[IE2_2]], i16 [[EE2_3]], i32 3
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE2_4:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 9
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE2_4:%.*]] = insertelement <8 x i16> [[IE2_3]], i16 [[EE2_4]], i32 4
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE2_5:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 11
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE2_5:%.*]] = insertelement <8 x i16> [[IE2_4]], i16 [[EE2_5]], i32 5
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE2_6:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 13
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[IE2_6:%.*]] = insertelement <8 x i16> [[IE2_5]], i16 [[EE2_6]], i32 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*VS.*[ ]*}} [[EE2_7:%.*]] = extractelement <16 x i16> [[LOAD2D]], i32 15
; CHECK: {{([0-9]+,[ ]*-512[ ]*).*VS.*[ ]*}} [[IE2_7:%.*]] = insertelement <8 x i16> [[IE2_6]], i16 [[EE2_7]], i32 7
; both DPAS increase the regpressure by 256 (sub vector of 256 dies, 512 created)
; CHECK: {{([0-9]+,[ ]*256[ ]*).*[ ]*}} [[DPAS1:%.*]] = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> [[IE1_7]], <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
; CHECK: {{([0-9]+,[ ]*256[ ]*).*[ ]*}} [[DPAS2:%.*]] = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> [[IE2_7]], <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
; CHECK: ret void
;
entry:
%base_addr = ptrtoint ptr addrspace(1) %A to i64
%load2d = call <16 x i16> @llvm.genx.GenISA.LSC2DBlockRead.v16i16(i64 %base_addr, i32 127, i32 1023, i32 127, i32 0, i32 0, i32 16, i32 16, i32 16, i32 2, i1 false, i1 false, i32 4)
%EE1.0 = extractelement <16 x i16> %load2d, i32 0
%IE1.0 = insertelement <8 x i16> undef, i16 %EE1.0, i32 0
%EE1.1 = extractelement <16 x i16> %load2d, i32 2
%IE1.1 = insertelement <8 x i16> %IE1.0, i16 %EE1.1, i32 1
%EE1.2 = extractelement <16 x i16> %load2d, i32 4
%IE1.2 = insertelement <8 x i16> %IE1.1, i16 %EE1.2, i32 2
%EE1.3 = extractelement <16 x i16> %load2d, i32 6
%IE1.3 = insertelement <8 x i16> %IE1.2, i16 %EE1.3, i32 3
%EE1.4 = extractelement <16 x i16> %load2d, i32 8
%IE1.4 = insertelement <8 x i16> %IE1.3, i16 %EE1.4, i32 4
%EE1.5 = extractelement <16 x i16> %load2d, i32 10
%IE1.5 = insertelement <8 x i16> %IE1.4, i16 %EE1.5, i32 5
%EE1.6 = extractelement <16 x i16> %load2d, i32 12
%IE1.6 = insertelement <8 x i16> %IE1.5, i16 %EE1.6, i32 6
%EE1.7 = extractelement <16 x i16> %load2d, i32 14
%IE1.7 = insertelement <8 x i16> %IE1.6, i16 %EE1.7, i32 7
%EE2.0 = extractelement <16 x i16> %load2d, i32 1
%IE2.0 = insertelement <8 x i16> undef, i16 %EE2.0, i32 0
%EE2.1 = extractelement <16 x i16> %load2d, i32 3
%IE2.1 = insertelement <8 x i16> %IE2.0, i16 %EE2.1, i32 1
%EE2.2 = extractelement <16 x i16> %load2d, i32 5
%IE2.2 = insertelement <8 x i16> %IE2.1, i16 %EE2.2, i32 2
%EE2.3 = extractelement <16 x i16> %load2d, i32 7
%IE2.3 = insertelement <8 x i16> %IE2.2, i16 %EE2.3, i32 3
%EE2.4 = extractelement <16 x i16> %load2d, i32 9
%IE2.4 = insertelement <8 x i16> %IE2.3, i16 %EE2.4, i32 4
%EE2.5 = extractelement <16 x i16> %load2d, i32 11
%IE2.5 = insertelement <8 x i16> %IE2.4, i16 %EE2.5, i32 5
%EE2.6 = extractelement <16 x i16> %load2d, i32 13
%IE2.6 = insertelement <8 x i16> %IE2.5, i16 %EE2.6, i32 6
%EE2.7 = extractelement <16 x i16> %load2d, i32 15
%IE2.7 = insertelement <8 x i16> %IE2.6, i16 %EE2.7, i32 7
%dpas1 = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> %IE1.7, <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
%dpas2 = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> %IE2.7, <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
ret void
}
define spir_kernel void @coalesced_scalars(ptr addrspace(1) %0) {
; CHECK: Function coalesced_scalars
; the IE instructions are marked as SCA. First IE adds regpressure
; then the last usage of the scalar (fadd) kills the hanging values
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*[ ]*}} [[TMP17:%.*]] = fmul fast float [[TMP9:%.*]], [[TMP1:%.*]]
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*[ ]*}} [[TMP18:%.*]] = fmul fast float [[TMP10:%.*]], [[TMP2:%.*]]
; CHECK: {{([0-9]+,[ ]*512[ ]*).*SCA.*[ ]*}} [[TMP19:%.*]] = insertelement <8 x float> zeroinitializer, float [[TMP17]], i64 0
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*[ ]*}} [[TMP20:%.*]] = fmul fast float [[TMP11:%.*]], [[TMP3:%.*]]
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*SCA.*[ ]*}} [[TMP21:%.*]] = insertelement <8 x float> [[TMP19]], float [[TMP18]], i64 1
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*[ ]*}} [[TMP22:%.*]] = fmul fast float [[TMP12:%.*]], [[TMP4:%.*]]
; CHECK: {{([0-9]+,[ ]*0[ ]*).*SCA.*[ ]*}} [[TMP23:%.*]] = insertelement <8 x float> [[TMP21]], float [[TMP20]], i64 2
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*[ ]*}} [[TMP24:%.*]] = fmul fast float [[TMP13:%.*]], [[TMP5:%.*]]
; CHECK: {{([0-9]+,[ ]*0[ ]*).*SCA.*[ ]*}} [[TMP25:%.*]] = insertelement <8 x float> [[TMP23]], float [[TMP22]], i64 3
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*[ ]*}} [[TMP26:%.*]] = fmul fast float [[TMP14:%.*]], [[TMP6:%.*]]
; CHECK: {{([0-9]+,[ ]*0[ ]*).*SCA.*[ ]*}} [[TMP27:%.*]] = insertelement <8 x float> [[TMP25]], float [[TMP24]], i64 4
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*[ ]*}} [[TMP28:%.*]] = fmul fast float [[TMP15:%.*]], [[TMP7:%.*]]
; CHECK: {{([0-9]+,[ ]*0[ ]*).*SCA.*[ ]*}} [[TMP29:%.*]] = insertelement <8 x float> [[TMP27]], float [[TMP26]], i64 5
; CHECK: {{([0-9]+,[ ]*[0-9-]+[ ]*).*[ ]*}} [[TMP30:%.*]] = fmul fast float [[TMP16:%.*]], [[TMP8:%.*]]
; CHECK: {{([0-9]+,[ ]*0[ ]*).*SCA.*[ ]*}} [[TMP31:%.*]] = insertelement <8 x float> [[TMP29]], float [[TMP28]], i64 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*SCA.*[ ]*}} [[TMP32:%.*]] = insertelement <8 x float> [[TMP31]], float [[TMP30]], i64 7
; dpas don't add any regpressure, they reuse the registers of the created vector
; CHECK: {{([0-9]+,[ ]*0[ ]*).*[ ]*}} [[TMP33:%.*]] = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> [[TMP32]], <8 x i16> zeroinitializer, <8 x i32> zeroinitializer, i32 0, i32 0, i32 0, i32 0, i1 false)
; CHECK: {{([0-9]+,[ ]*0[ ]*).*[ ]*}} [[TMP34:%.*]] = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> [[TMP33]], <8 x i16> zeroinitializer, <8 x i32> zeroinitializer, i32 0, i32 0, i32 0, i32 0, i1 false)
; extract the values from the vector. The extractelement instructions are marked as V2S and don't increase regpressure
; the vector hangs, so the last EE doesn't reduce regpressure. In this case the EEs are used in PHI nodes so these hanging vals are not killed
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[TMP35:%.*]] = extractelement <8 x float> [[TMP34]], i64 0
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[TMP36:%.*]] = extractelement <8 x float> [[TMP34]], i64 1
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[TMP37:%.*]] = extractelement <8 x float> [[TMP34]], i64 2
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[TMP38:%.*]] = extractelement <8 x float> [[TMP34]], i64 3
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[TMP39:%.*]] = extractelement <8 x float> [[TMP34]], i64 4
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[TMP40:%.*]] = extractelement <8 x float> [[TMP34]], i64 5
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[TMP41:%.*]] = extractelement <8 x float> [[TMP34]], i64 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[TMP42:%.*]] = extractelement <8 x float> [[TMP34]], i64 7
; fadd kills the hanging values from SCA and creates one float -> -448
; CHECK: {{([0-9]+,[ ]*-448[ ]*).*[ ]*}} [[TMP50:%.*]] = fadd fast float [[TMP37:%.*]], [[TMP26]]
; this fadd doesn't kill the hanging EEs (they are used in PHI nodes), so it increases regpressure by 64
; CHECK: {{([0-9]+,[ ]*64[ ]*).*[ ]*}} [[TMP51:%.*]] = fadd fast float [[TMP42:%.*]], 4.000000e+00
entry:
br label %._crit_edge
._crit_edge: ; preds = %._crit_edge, %0
%1 = phi float [ 0.000000e+00, %entry ], [ %35, %._crit_edge ]
%2 = phi float [ 0.000000e+00, %entry ], [ %36, %._crit_edge ]
%3 = phi float [ 0.000000e+00, %entry ], [ %37, %._crit_edge ]
%4 = phi float [ 0.000000e+00, %entry ], [ %38, %._crit_edge ]
%5 = phi float [ 0.000000e+00, %entry ], [ %39, %._crit_edge ]
%6 = phi float [ 0.000000e+00, %entry ], [ %40, %._crit_edge ]
%7 = phi float [ 0.000000e+00, %entry ], [ %41, %._crit_edge ]
%8 = phi float [ 0.000000e+00, %entry ], [ %42, %._crit_edge ]
%9 = call float @llvm.exp2.f32(float 0.000000e+00)
%10 = call float @llvm.exp2.f32(float 0.000000e+00)
%11 = call float @llvm.exp2.f32(float 0.000000e+00)
%12 = call float @llvm.exp2.f32(float 0.000000e+00)
%13 = call float @llvm.exp2.f32(float 0.000000e+00)
%14 = call float @llvm.exp2.f32(float 0.000000e+00)
%15 = call float @llvm.exp2.f32(float 0.000000e+00)
%16 = call float @llvm.exp2.f32(float 0.000000e+00)
%17 = fmul fast float %9, %1
%18 = fmul fast float %10, %2
%19 = fmul fast float %11, %3
%20 = fmul fast float %12, %4
%21 = fmul fast float %13, %5
%22 = fmul fast float %14, %6
%23 = fmul fast float %15, %7
%24 = fmul fast float %16, %8
%25 = insertelement <8 x float> zeroinitializer, float %17, i64 0
%26 = insertelement <8 x float> %25, float %18, i64 1
%27 = insertelement <8 x float> %26, float %19, i64 2
%28 = insertelement <8 x float> %27, float %20, i64 3
%29 = insertelement <8 x float> %28, float %21, i64 4
%30 = insertelement <8 x float> %29, float %22, i64 5
%31 = insertelement <8 x float> %30, float %23, i64 6
%32 = insertelement <8 x float> %31, float %24, i64 7
%33 = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> %32, <8 x i16> zeroinitializer, <8 x i32> zeroinitializer, i32 0, i32 0, i32 0, i32 0, i1 false)
%34 = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> %33, <8 x i16> zeroinitializer, <8 x i32> zeroinitializer, i32 0, i32 0, i32 0, i32 0, i1 false)
%35 = extractelement <8 x float> %34, i64 0
%36 = extractelement <8 x float> %34, i64 1
%37 = extractelement <8 x float> %34, i64 2
%38 = extractelement <8 x float> %34, i64 3
%39 = extractelement <8 x float> %34, i64 4
%40 = extractelement <8 x float> %34, i64 5
%41 = extractelement <8 x float> %34, i64 6
%42 = extractelement <8 x float> %34, i64 7
%43 = fadd fast float %37, %22
%44 = fadd fast float %42, 4.000000e+00
br label %._crit_edge
}
define spir_kernel void @vector_to_scalars_pattern(ptr addrspace(1) %A) {
; CHECK: Function vector_to_scalars_pattern
; DPAS increases regpressure
; CHECK: {{([0-9]+,[ ]*512[ ]*).*[ ]*}} [[DPAS:%.*]] = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> undef, <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
; EE don't increase regpressure
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[EE0:%.*]] = extractelement <8 x float> [[DPAS]], i64 0
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[EE1:%.*]] = extractelement <8 x float> [[DPAS]], i64 1
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[EE2:%.*]] = extractelement <8 x float> [[DPAS]], i64 2
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[EE3:%.*]] = extractelement <8 x float> [[DPAS]], i64 3
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[EE4:%.*]] = extractelement <8 x float> [[DPAS]], i64 4
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[EE5:%.*]] = extractelement <8 x float> [[DPAS]], i64 5
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[EE6:%.*]] = extractelement <8 x float> [[DPAS]], i64 6
; CHECK: {{([0-9]+,[ ]*0[ ]*).*V2S.*[ ]*}} [[EE7:%.*]] = extractelement <8 x float> [[DPAS]], i64 7
; The vector doesn't die on the last EE, it hangs.
; The uses of the EEs increase regpressure
; CHECK: {{([0-9]+,[ ]*64[ ]*).*[ ]*}} [[USE0:%.*]] = fadd fast float [[EE0]], 1.0
; CHECK: {{([0-9]+,[ ]*64[ ]*).*[ ]*}} [[USE1:%.*]] = fadd fast float [[EE1]], 2.0
; CHECK: {{([0-9]+,[ ]*64[ ]*).*[ ]*}} [[USE2:%.*]] = fadd fast float [[EE2]], 3.0
; CHECK: {{([0-9]+,[ ]*64[ ]*).*[ ]*}} [[USE3:%.*]] = fadd fast float [[EE3]], 4.0
; CHECK: {{([0-9]+,[ ]*64[ ]*).*[ ]*}} [[USE4:%.*]] = fadd fast float [[EE4]], 5.0
; CHECK: {{([0-9]+,[ ]*64[ ]*).*[ ]*}} [[USE5:%.*]] = fadd fast float [[EE5]], 6.0
; CHECK: {{([0-9]+,[ ]*64[ ]*).*[ ]*}} [[USE6:%.*]] = fadd fast float [[EE6]], 7.0
; The vector dies on the last EE usage
; CHECK: {{([0-9]+,[ ]*-448[ ]*).*[ ]*}} [[USE7:%.*]] = fadd fast float [[EE7]], 8.0
; CHECK: ret void
entry:
%dpas = call <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(<8 x float> zeroinitializer, <8 x i16> undef, <8 x i32> zeroinitializer, i32 1, i32 1, i32 1, i32 1, i1 false)
%ee0 = extractelement <8 x float> %dpas, i64 0
%ee1 = extractelement <8 x float> %dpas, i64 1
%ee2 = extractelement <8 x float> %dpas, i64 2
%ee3 = extractelement <8 x float> %dpas, i64 3
%ee4 = extractelement <8 x float> %dpas, i64 4
%ee5 = extractelement <8 x float> %dpas, i64 5
%ee6 = extractelement <8 x float> %dpas, i64 6
%ee7 = extractelement <8 x float> %dpas, i64 7
%use0 = fadd fast float %ee0, 1.0
%use1 = fadd fast float %ee1, 2.0
%use2 = fadd fast float %ee2, 3.0
%use3 = fadd fast float %ee3, 4.0
%use4 = fadd fast float %ee4, 5.0
%use5 = fadd fast float %ee5, 6.0
%use6 = fadd fast float %ee6, 7.0
%use7 = fadd fast float %ee7, 8.0
ret void
}
declare <8 x float> @llvm.genx.GenISA.sub.group.dpas.v8f32.v8f32.v8i16.v8i32(
<8 x float>, <8 x i16>, <8 x i32>, i32, i32, i32, i32, i1) #1
declare float @llvm.exp2.f32(float) #2
declare <16 x i16> @llvm.genx.GenISA.LSC2DBlockRead.v16i16(i64, i32, i32, i32, i32, i32, i32, i32, i32, i32, i1, i1, i32) #1
attributes #0 = { convergent nounwind }
attributes #1 = { convergent nounwind readnone willreturn }
attributes #2 = { nofree nosync nounwind readnone speculatable willreturn }
|
