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
|
;=========================== begin_copyright_notice ============================
;
; Copyright (C) 2023 Intel Corporation
;
; SPDX-License-Identifier: MIT
;
;============================ end_copyright_notice =============================
; RUN: %opt %use_old_pass_manager% -enable-debugify -GenXLegacyToLscTranslator -march=genx64 -mcpu=XeHPC -mtriple=spir64-unknown-unknown -mattr=+translate_legacy_message -S < %s 2>&1 | FileCheck %s
target datalayout = "e-p:64:64-i64:64-n8:16:32:64"
target triple = "genx64-unknown-unknown"
declare <16 x i32> @llvm.genx.gather.scaled.v16i32.v16i1.v16i32(<16 x i1>, i32, i16, i32, i32, <16 x i32>, <16 x i32>)
declare <16 x i32> @llvm.genx.gather.scaled2.v16i32.v16i1.v16i32(i32, i16, i32, i32, <16 x i32>)
declare <16 x i32> @llvm.genx.gather.masked.scaled2.v16i32.v16i32.v16i1(i32, i16, i32, i32, <16 x i32>, <16 x i1>)
declare void @llvm.genx.scatter.scaled.v16i1.v16i32.v16i32(<16 x i1>, i32, i16, i32, i32, <16 x i32>, <16 x i32>)
declare <64 x i8> @llvm.genx.svm.gather.v64i8.v16i1.v16i64(<16 x i1>, i32, <16 x i64>, <64 x i8>)
declare <128 x i8> @llvm.genx.svm.gather.v128i8.v16i1.v16i64(<16 x i1>, i32, <16 x i64>, <128 x i8>)
declare void @llvm.genx.svm.scatter.v16i1.v16i64.v64i8(<16 x i1>, i32, <16 x i64>, <64 x i8>)
declare void @llvm.genx.svm.scatter.v16i1.v16i64.v128i8(<16 x i1>, i32, <16 x i64>, <128 x i8>)
declare <16 x i32> @llvm.genx.svm.gather.v16i32.v16i1.v16i64(<16 x i1>, i32, <16 x i64>, <16 x i32>)
declare <32 x i32> @llvm.genx.svm.gather.v32i32.v16i1.v16i64(<16 x i1>, i32, <16 x i64>, <32 x i32>)
declare <64 x i32> @llvm.genx.svm.gather.v64i32.v16i1.v16i64(<16 x i1>, i32, <16 x i64>, <64 x i32>)
declare <16 x i64> @llvm.genx.svm.gather.v16i64.v16i1.v16i64(<16 x i1>, i32, <16 x i64>, <16 x i64>)
declare <32 x i64> @llvm.genx.svm.gather.v32i64.v16i1.v16i64(<16 x i1>, i32, <16 x i64>, <32 x i64>)
declare <64 x i64> @llvm.genx.svm.gather.v64i64.v16i1.v16i64(<16 x i1>, i32, <16 x i64>, <64 x i64>)
declare void @llvm.genx.svm.scatter.v16i1.v16i64.v16i32(<16 x i1>, i32, <16 x i64>, <16 x i32>)
declare void @llvm.genx.svm.scatter.v16i1.v16i64.v32i32(<16 x i1>, i32, <16 x i64>, <32 x i32>)
declare void @llvm.genx.svm.scatter.v16i1.v16i64.v64i32(<16 x i1>, i32, <16 x i64>, <64 x i32>)
declare void @llvm.genx.svm.scatter.v16i1.v16i64.v16i64(<16 x i1>, i32, <16 x i64>, <16 x i64>)
declare void @llvm.genx.svm.scatter.v16i1.v16i64.v32i64(<16 x i1>, i32, <16 x i64>, <32 x i64>)
declare void @llvm.genx.svm.scatter.v16i1.v16i64.v64i64(<16 x i1>, i32, <16 x i64>, <64 x i64>)
; CHECK-LABEL: test_bti
define void @test_bti(<16 x i1> %pred, i32 %offset, <16 x i32> %addr, <16 x i32> %data) {
; CHECK: [[INS1:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT1:%[^ ]+]] = shufflevector <16 x i32> [[INS1]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR1:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT1]]
; CHECK: %ld8 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 5, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR1]], i16 1, i32 0, <16 x i32> %data)
%ld8 = call <16 x i32> @llvm.genx.gather.scaled.v16i32.v16i1.v16i32(<16 x i1> %pred, i32 0, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS2:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT2:%[^ ]+]] = shufflevector <16 x i32> [[INS2]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR2:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT2]]
; CHECK: %ld16 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 6, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR2]], i16 1, i32 0, <16 x i32> %data)
%ld16 = call <16 x i32> @llvm.genx.gather.scaled.v16i32.v16i1.v16i32(<16 x i1> %pred, i32 1, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS3:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT3:%[^ ]+]] = shufflevector <16 x i32> [[INS3]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR3:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT3]]
; CHECK: %ld32 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 3, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR3]], i16 1, i32 0, <16 x i32> %data)
%ld32 = call <16 x i32> @llvm.genx.gather.scaled.v16i32.v16i1.v16i32(<16 x i1> %pred, i32 2, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS4:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT4:%[^ ]+]] = shufflevector <16 x i32> [[INS4]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR4:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT4]]
; CHECK: %ldx8 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> {{<(i1 true(, )?){16}>}}, i8 2, i8 5, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR4]], i16 1, i32 0, <16 x i32> undef)
%ldx8 = call <16 x i32> @llvm.genx.gather.scaled2.v16i32.v16i1.v16i32(i32 0, i16 0, i32 1, i32 %offset, <16 x i32> %addr)
; CHECK: [[INS5:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT5:%[^ ]+]] = shufflevector <16 x i32> [[INS5]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR5:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT5]]
; CHECK: %ldx16 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> {{<(i1 true(, )?){16}>}}, i8 2, i8 6, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR5]], i16 1, i32 0, <16 x i32> undef)
%ldx16 = call <16 x i32> @llvm.genx.gather.scaled2.v16i32.v16i1.v16i32(i32 1, i16 0, i32 1, i32 %offset, <16 x i32> %addr)
; CHECK: [[INS6:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT6:%[^ ]+]] = shufflevector <16 x i32> [[INS6]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR6:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT6]]
; CHECK: %ldx32 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> {{<(i1 true(, )?){16}>}}, i8 2, i8 3, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR6]], i16 1, i32 0, <16 x i32> undef)
%ldx32 = call <16 x i32> @llvm.genx.gather.scaled2.v16i32.v16i1.v16i32(i32 2, i16 0, i32 1, i32 %offset, <16 x i32> %addr)
; CHECK: [[INS7:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT7:%[^ ]+]] = shufflevector <16 x i32> [[INS7]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR7:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT7]]
; CHECK: %ldm8 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 5, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR7]], i16 1, i32 0, <16 x i32> undef)
%ldm8 = call <16 x i32> @llvm.genx.gather.masked.scaled2.v16i32.v16i32.v16i1(i32 0, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i1> %pred)
; CHECK: [[INS8:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT8:%[^ ]+]] = shufflevector <16 x i32> [[INS8]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR8:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT8]]
; CHECK: %ldm16 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 6, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR8]], i16 1, i32 0, <16 x i32> undef)
%ldm16 = call <16 x i32> @llvm.genx.gather.masked.scaled2.v16i32.v16i32.v16i1(i32 1, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i1> %pred)
; CHECK: [[INS9:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT9:%[^ ]+]] = shufflevector <16 x i32> [[INS9]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR9:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT9]]
; CHECK: %ldm32 = call <16 x i32> @llvm.vc.internal.lsc.load.bti.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 3, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR9]], i16 1, i32 0, <16 x i32> undef)
%ldm32 = call <16 x i32> @llvm.genx.gather.masked.scaled2.v16i32.v16i32.v16i1(i32 2, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i1> %pred)
; CHECK: [[INS10:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT10:%[^ ]+]] = shufflevector <16 x i32> [[INS10]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR10:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT10]]
; CHECK: call void @llvm.vc.internal.lsc.store.bti.v16i1.v2i8.v16i32.v16i32(<16 x i1> %pred, i8 2, i8 5, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR10]], i16 1, i32 0, <16 x i32> %data)
call void @llvm.genx.scatter.scaled.v16i1.v16i32.v16i32(<16 x i1> %pred, i32 0, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS11:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT11:%[^ ]+]] = shufflevector <16 x i32> [[INS11]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR11:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT11]]
; CHECK: call void @llvm.vc.internal.lsc.store.bti.v16i1.v2i8.v16i32.v16i32(<16 x i1> %pred, i8 2, i8 6, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR11]], i16 1, i32 0, <16 x i32> %data)
call void @llvm.genx.scatter.scaled.v16i1.v16i32.v16i32(<16 x i1> %pred, i32 1, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS12:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT12:%[^ ]+]] = shufflevector <16 x i32> [[INS12]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR12:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT12]]
; CHECK: call void @llvm.vc.internal.lsc.store.bti.v16i1.v2i8.v16i32.v16i32(<16 x i1> %pred, i8 2, i8 3, i8 1, <2 x i8> zeroinitializer, i32 1, <16 x i32> [[ADDR12]], i16 1, i32 0, <16 x i32> %data)
call void @llvm.genx.scatter.scaled.v16i1.v16i32.v16i32(<16 x i1> %pred, i32 2, i16 0, i32 1, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
ret void
}
; CHECK-LABEL: test_slm
define void @test_slm(<16 x i1> %pred, i32 %offset, <16 x i32> %addr, <16 x i32> %data) {
; CHECK: [[INS1:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT1:%[^ ]+]] = shufflevector <16 x i32> [[INS1]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR1:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT1]]
; CHECK: %ld8 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 5, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR1]], i16 1, i32 0, <16 x i32> %data)
%ld8 = call <16 x i32> @llvm.genx.gather.scaled.v16i32.v16i1.v16i32(<16 x i1> %pred, i32 0, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS2:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT2:%[^ ]+]] = shufflevector <16 x i32> [[INS2]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR2:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT2]]
; CHECK: %ld16 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 6, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR2]], i16 1, i32 0, <16 x i32> %data)
%ld16 = call <16 x i32> @llvm.genx.gather.scaled.v16i32.v16i1.v16i32(<16 x i1> %pred, i32 1, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS3:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT3:%[^ ]+]] = shufflevector <16 x i32> [[INS3]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR3:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT3]]
; CHECK: %ld32 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 3, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR3]], i16 1, i32 0, <16 x i32> %data)
%ld32 = call <16 x i32> @llvm.genx.gather.scaled.v16i32.v16i1.v16i32(<16 x i1> %pred, i32 2, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS4:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT4:%[^ ]+]] = shufflevector <16 x i32> [[INS4]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR4:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT4]]
; CHECK: %ldx8 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> {{<(i1 true(, )?){16}>}}, i8 2, i8 5, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR4]], i16 1, i32 0, <16 x i32> undef)
%ldx8 = call <16 x i32> @llvm.genx.gather.scaled2.v16i32.v16i1.v16i32(i32 0, i16 0, i32 254, i32 %offset, <16 x i32> %addr)
; CHECK: [[INS5:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT5:%[^ ]+]] = shufflevector <16 x i32> [[INS5]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR5:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT5]]
; CHECK: %ldx16 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> {{<(i1 true(, )?){16}>}}, i8 2, i8 6, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR5]], i16 1, i32 0, <16 x i32> undef)
%ldx16 = call <16 x i32> @llvm.genx.gather.scaled2.v16i32.v16i1.v16i32(i32 1, i16 0, i32 254, i32 %offset, <16 x i32> %addr)
; CHECK: [[INS6:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT6:%[^ ]+]] = shufflevector <16 x i32> [[INS6]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR6:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT6]]
; CHECK: %ldx32 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> {{<(i1 true(, )?){16}>}}, i8 2, i8 3, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR6]], i16 1, i32 0, <16 x i32> undef)
%ldx32 = call <16 x i32> @llvm.genx.gather.scaled2.v16i32.v16i1.v16i32(i32 2, i16 0, i32 254, i32 %offset, <16 x i32> %addr)
; CHECK: [[INS7:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT7:%[^ ]+]] = shufflevector <16 x i32> [[INS7]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR7:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT7]]
; CHECK: %ldm8 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 5, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR7]], i16 1, i32 0, <16 x i32> undef)
%ldm8 = call <16 x i32> @llvm.genx.gather.masked.scaled2.v16i32.v16i32.v16i1(i32 0, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i1> %pred)
; CHECK: [[INS8:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT8:%[^ ]+]] = shufflevector <16 x i32> [[INS8]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR8:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT8]]
; CHECK: %ldm16 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 6, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR8]], i16 1, i32 0, <16 x i32> undef)
%ldm16 = call <16 x i32> @llvm.genx.gather.masked.scaled2.v16i32.v16i32.v16i1(i32 1, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i1> %pred)
; CHECK: [[INS9:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT9:%[^ ]+]] = shufflevector <16 x i32> [[INS9]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR9:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT9]]
; CHECK: %ldm32 = call <16 x i32> @llvm.vc.internal.lsc.load.slm.v16i32.v16i1.v2i8.v16i32(<16 x i1> %pred, i8 2, i8 3, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR9]], i16 1, i32 0, <16 x i32> undef)
%ldm32 = call <16 x i32> @llvm.genx.gather.masked.scaled2.v16i32.v16i32.v16i1(i32 2, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i1> %pred)
; CHECK: [[INS10:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT10:%[^ ]+]] = shufflevector <16 x i32> [[INS10]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR10:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT10]]
; CHECK: call void @llvm.vc.internal.lsc.store.slm.v16i1.v2i8.v16i32.v16i32(<16 x i1> %pred, i8 2, i8 5, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR10]], i16 1, i32 0, <16 x i32> %data)
call void @llvm.genx.scatter.scaled.v16i1.v16i32.v16i32(<16 x i1> %pred, i32 0, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS11:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT11:%[^ ]+]] = shufflevector <16 x i32> [[INS11]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR11:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT11]]
; CHECK: call void @llvm.vc.internal.lsc.store.slm.v16i1.v2i8.v16i32.v16i32(<16 x i1> %pred, i8 2, i8 6, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR11]], i16 1, i32 0, <16 x i32> %data)
call void @llvm.genx.scatter.scaled.v16i1.v16i32.v16i32(<16 x i1> %pred, i32 1, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
; CHECK: [[INS12:%[^ ]+]] = insertelement <16 x i32> poison, i32 %offset, i32 0
; CHECK: [[SPLAT12:%[^ ]+]] = shufflevector <16 x i32> [[INS12]], <16 x i32> poison, <16 x i32> zeroinitializer
; CHECK: [[ADDR12:%[^ ]+]] = add <16 x i32> %addr, [[SPLAT12]]
; CHECK: call void @llvm.vc.internal.lsc.store.slm.v16i1.v2i8.v16i32.v16i32(<16 x i1> %pred, i8 2, i8 3, i8 1, <2 x i8> zeroinitializer, i32 0, <16 x i32> [[ADDR12]], i16 1, i32 0, <16 x i32> %data)
call void @llvm.genx.scatter.scaled.v16i1.v16i32.v16i32(<16 x i1> %pred, i32 2, i16 0, i32 254, i32 %offset, <16 x i32> %addr, <16 x i32> %data)
ret void
}
; CHECK-LABEL: test_ugm_unaligned_dword
define void @test_ugm_unaligned_dword(<16 x i1> %pred, <16 x i64> %addr, <64 x i8> %data) {
; CHECK: [[PRECAST1:%[^ ]+]] = bitcast <64 x i8> %data to <16 x i32>
; CHECK: [[LD1:%[^ ]+]] = call <16 x i32> @llvm.vc.internal.lsc.load.ugm.v16i32.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 5, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i32> [[PRECAST1]])
; CHECK: %ld8 = bitcast <16 x i32> [[LD1]] to <64 x i8>
%ld8 = call <64 x i8> @llvm.genx.svm.gather.v64i8.v16i1.v16i64(<16 x i1> %pred, i32 0, <16 x i64> %addr, <64 x i8> %data)
; CHECK: [[PRECAST2:%[^ ]+]] = bitcast <64 x i8> %data to <16 x i32>
; CHECK: [[LD2:%[^ ]+]] = call <16 x i32> @llvm.vc.internal.lsc.load.ugm.v16i32.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 6, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i32> [[PRECAST2]])
; CHECK: %ld16 = bitcast <16 x i32> [[LD2]] to <64 x i8>
%ld16 = call <64 x i8> @llvm.genx.svm.gather.v64i8.v16i1.v16i64(<16 x i1> %pred, i32 1, <16 x i64> %addr, <64 x i8> %data)
; CHECK: [[PRECAST3:%[^ ]+]] = bitcast <64 x i8> %data to <16 x i32>
; CHECK: [[LD3:%[^ ]+]] = call <16 x i32> @llvm.vc.internal.lsc.load.ugm.v16i32.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 3, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i32> [[PRECAST3]])
; CHECK: %ld32 = bitcast <16 x i32> [[LD3]] to <64 x i8>
%ld32 = call <64 x i8> @llvm.genx.svm.gather.v64i8.v16i1.v16i64(<16 x i1> %pred, i32 2, <16 x i64> %addr, <64 x i8> %data)
; CHECK: [[PRECAST4:%[^ ]+]] = bitcast <64 x i8> %data to <16 x i32>
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v16i32(<16 x i1> %pred, i8 3, i8 5, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i32> [[PRECAST4]])
call void @llvm.genx.svm.scatter.v16i1.v16i64.v64i8(<16 x i1> %pred, i32 0, <16 x i64> %addr, <64 x i8> %data)
; CHECK: [[PRECAST5:%[^ ]+]] = bitcast <64 x i8> %data to <16 x i32>
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v16i32(<16 x i1> %pred, i8 3, i8 6, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i32> [[PRECAST5]])
call void @llvm.genx.svm.scatter.v16i1.v16i64.v64i8(<16 x i1> %pred, i32 1, <16 x i64> %addr, <64 x i8> %data)
; CHECK: [[PRECAST6:%[^ ]+]] = bitcast <64 x i8> %data to <16 x i32>
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v16i32(<16 x i1> %pred, i8 3, i8 3, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i32> [[PRECAST6]])
call void @llvm.genx.svm.scatter.v16i1.v16i64.v64i8(<16 x i1> %pred, i32 2, <16 x i64> %addr, <64 x i8> %data)
ret void
}
; CHECK-LABEL: test_ugm_unaligned_oword
define void @test_ugm_unaligned_oword(<16 x i1> %pred, <16 x i64> %addr, <128 x i8> %data) {
; CHECK: [[PRECAST1:%[^ ]+]] = bitcast <128 x i8> %data to <16 x i64>
; CHECK: [[LD1:%[^ ]+]] = call <16 x i64> @llvm.vc.internal.lsc.load.ugm.v16i64.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 4, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i64> [[PRECAST1]])
; CHECK: %ld64 = bitcast <16 x i64> [[LD1]] to <128 x i8>
%ld64 = call <128 x i8> @llvm.genx.svm.gather.v128i8.v16i1.v16i64(<16 x i1> %pred, i32 3, <16 x i64> %addr, <128 x i8> %data)
; CHECK: [[PRECAST6:%[^ ]+]] = bitcast <128 x i8> %data to <16 x i64>
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v16i64(<16 x i1> %pred, i8 3, i8 4, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i64> [[PRECAST6]])
call void @llvm.genx.svm.scatter.v16i1.v16i64.v128i8(<16 x i1> %pred, i32 3, <16 x i64> %addr, <128 x i8> %data)
ret void
}
; CHECK-LABEL: test_ugm
define void @test_ugm(<16 x i1> %pred, <16 x i64> %addr) {
; CHECK: %ld1 = call <16 x i32> @llvm.vc.internal.lsc.load.ugm.v16i32.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 3, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i32> undef)
; CHECK: %ld2 = call <32 x i32> @llvm.vc.internal.lsc.load.ugm.v32i32.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 3, i8 2, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <32 x i32> undef)
; CHECK: %ld4 = call <64 x i32> @llvm.vc.internal.lsc.load.ugm.v64i32.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 3, i8 4, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <64 x i32> undef)
%ld1 = call <16 x i32> @llvm.genx.svm.gather.v16i32.v16i1.v16i64(<16 x i1> %pred, i32 0, <16 x i64> %addr, <16 x i32> undef)
%ld2 = call <32 x i32> @llvm.genx.svm.gather.v32i32.v16i1.v16i64(<16 x i1> %pred, i32 1, <16 x i64> %addr, <32 x i32> undef)
%ld4 = call <64 x i32> @llvm.genx.svm.gather.v64i32.v16i1.v16i64(<16 x i1> %pred, i32 2, <16 x i64> %addr, <64 x i32> undef)
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v16i32(<16 x i1> %pred, i8 3, i8 3, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i32> %ld1)
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v32i32(<16 x i1> %pred, i8 3, i8 3, i8 2, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <32 x i32> %ld2)
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v64i32(<16 x i1> %pred, i8 3, i8 3, i8 4, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <64 x i32> %ld4)
call void @llvm.genx.svm.scatter.v16i1.v16i64.v16i32(<16 x i1> %pred, i32 0, <16 x i64> %addr, <16 x i32> %ld1)
call void @llvm.genx.svm.scatter.v16i1.v16i64.v32i32(<16 x i1> %pred, i32 1, <16 x i64> %addr, <32 x i32> %ld2)
call void @llvm.genx.svm.scatter.v16i1.v16i64.v64i32(<16 x i1> %pred, i32 2, <16 x i64> %addr, <64 x i32> %ld4)
; CHECK: %lq1 = call <16 x i64> @llvm.vc.internal.lsc.load.ugm.v16i64.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 4, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i64> undef)
; CHECK: %lq2 = call <32 x i64> @llvm.vc.internal.lsc.load.ugm.v32i64.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 4, i8 2, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <32 x i64> undef)
; CHECK: %lq4 = call <64 x i64> @llvm.vc.internal.lsc.load.ugm.v64i64.v16i1.v2i8.v16i64(<16 x i1> %pred, i8 3, i8 4, i8 4, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <64 x i64> undef)
%lq1 = call <16 x i64> @llvm.genx.svm.gather.v16i64.v16i1.v16i64(<16 x i1> %pred, i32 0, <16 x i64> %addr, <16 x i64> undef)
%lq2 = call <32 x i64> @llvm.genx.svm.gather.v32i64.v16i1.v16i64(<16 x i1> %pred, i32 1, <16 x i64> %addr, <32 x i64> undef)
%lq4 = call <64 x i64> @llvm.genx.svm.gather.v64i64.v16i1.v16i64(<16 x i1> %pred, i32 2, <16 x i64> %addr, <64 x i64> undef)
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v16i64(<16 x i1> %pred, i8 3, i8 4, i8 1, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <16 x i64> %lq1)
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v32i64(<16 x i1> %pred, i8 3, i8 4, i8 2, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <32 x i64> %lq2)
; CHECK: call void @llvm.vc.internal.lsc.store.ugm.v16i1.v2i8.v16i64.v64i64(<16 x i1> %pred, i8 3, i8 4, i8 4, <2 x i8> zeroinitializer, i64 0, <16 x i64> %addr, i16 1, i32 0, <64 x i64> %lq4)
call void @llvm.genx.svm.scatter.v16i1.v16i64.v16i64(<16 x i1> %pred, i32 0, <16 x i64> %addr, <16 x i64> %lq1)
call void @llvm.genx.svm.scatter.v16i1.v16i64.v32i64(<16 x i1> %pred, i32 1, <16 x i64> %addr, <32 x i64> %lq2)
call void @llvm.genx.svm.scatter.v16i1.v16i64.v64i64(<16 x i1> %pred, i32 2, <16 x i64> %addr, <64 x i64> %lq4)
ret void
}
|
