File: promote-kernel-arguments.ll

package info (click to toggle)
llvm-toolchain-14 1%3A14.0.6-12
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 1,496,180 kB
  • sloc: cpp: 5,593,972; ansic: 986,872; asm: 585,869; python: 184,223; objc: 72,530; lisp: 31,119; f90: 27,793; javascript: 9,780; pascal: 9,762; sh: 9,482; perl: 7,468; ml: 5,432; awk: 3,523; makefile: 2,538; xml: 953; cs: 573; fortran: 567
file content (317 lines) | stat: -rw-r--r-- 19,316 bytes parent folder | download | duplicates (3) 
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
 
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S -mtriple=amdgcn-amd-amdhsa < %s -amdgpu-promote-kernel-arguments -infer-address-spaces | FileCheck %s
; RUN: opt -S -mtriple=amdgcn-amd-amdhsa < %s -passes=amdgpu-promote-kernel-arguments,infer-address-spaces | FileCheck %s
; RUN: opt -S -mtriple=amdgcn-amd-amdhsa < %s -amdgpu-promote-kernel-arguments -infer-address-spaces | llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 | FileCheck -check-prefix=GCN %s

; GCN-LABEL: ptr_nest_3:
; GCN-COUNT-2: global_load_dwordx2
; GCN:         global_store_dword
define amdgpu_kernel void @ptr_nest_3(float** addrspace(1)* nocapture readonly %Arg) {
; CHECK-LABEL: @ptr_nest_3(
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[I:%.*]] = tail call i32 @llvm.amdgcn.workitem.id.x()
; CHECK-NEXT:    [[P1:%.*]] = getelementptr inbounds float**, float** addrspace(1)* [[ARG:%.*]], i32 [[I]]
; CHECK-NEXT:    [[P2:%.*]] = load float**, float** addrspace(1)* [[P1]], align 8
; CHECK-NEXT:    [[P2_GLOBAL:%.*]] = addrspacecast float** [[P2]] to float* addrspace(1)*
; CHECK-NEXT:    [[P3:%.*]] = load float*, float* addrspace(1)* [[P2_GLOBAL]], align 8
; CHECK-NEXT:    [[P3_GLOBAL:%.*]] = addrspacecast float* [[P3]] to float addrspace(1)*
; CHECK-NEXT:    store float 0.000000e+00, float addrspace(1)* [[P3_GLOBAL]], align 4
; CHECK-NEXT:    ret void
;
entry:
  %i = tail call i32 @llvm.amdgcn.workitem.id.x()
  %p1 = getelementptr inbounds float**, float** addrspace(1)* %Arg, i32 %i
  %p2 = load float**, float** addrspace(1)* %p1, align 8
  %p3 = load float*, float** %p2, align 8
  store float 0.000000e+00, float* %p3, align 4
  ret void
}

; GCN-LABEL: ptr_bitcast:
; GCN: global_load_dwordx2
; GCN: global_store_dword
define amdgpu_kernel void @ptr_bitcast(float** nocapture readonly %Arg) {
; CHECK-LABEL: @ptr_bitcast(
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[ARG_GLOBAL:%.*]] = addrspacecast float** [[ARG:%.*]] to float* addrspace(1)*
; CHECK-NEXT:    [[I:%.*]] = tail call i32 @llvm.amdgcn.workitem.id.x()
; CHECK-NEXT:    [[P1:%.*]] = getelementptr inbounds float*, float* addrspace(1)* [[ARG_GLOBAL]], i32 [[I]]
; CHECK-NEXT:    [[P1_CAST:%.*]] = bitcast float* addrspace(1)* [[P1]] to i32* addrspace(1)*
; CHECK-NEXT:    [[P2:%.*]] = load i32*, i32* addrspace(1)* [[P1_CAST]], align 8
; CHECK-NEXT:    [[P2_GLOBAL:%.*]] = addrspacecast i32* [[P2]] to i32 addrspace(1)*
; CHECK-NEXT:    store i32 0, i32 addrspace(1)* [[P2_GLOBAL]], align 4
; CHECK-NEXT:    ret void
;
entry:
  %i = tail call i32 @llvm.amdgcn.workitem.id.x()
  %p1 = getelementptr inbounds float*, float** %Arg, i32 %i
  %p1.cast = bitcast float** %p1 to i32**
  %p2 = load i32*, i32** %p1.cast, align 8
  store i32 0, i32* %p2, align 4
  ret void
}

%struct.S = type { float* }

; GCN-LABEL: ptr_in_struct:
; GCN: s_load_dwordx2
; GCN: global_store_dword
define amdgpu_kernel void @ptr_in_struct(%struct.S addrspace(1)* nocapture readonly %Arg) {
; CHECK-LABEL: @ptr_in_struct(
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[P:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], [[STRUCT_S]] addrspace(1)* [[ARG:%.*]], i64 0, i32 0
; CHECK-NEXT:    [[P1:%.*]] = load float*, float* addrspace(1)* [[P]], align 8
; CHECK-NEXT:    [[P1_GLOBAL:%.*]] = addrspacecast float* [[P1]] to float addrspace(1)*
; CHECK-NEXT:    [[ID:%.*]] = tail call i32 @llvm.amdgcn.workitem.id.x()
; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, float addrspace(1)* [[P1_GLOBAL]], i32 [[ID]]
; CHECK-NEXT:    store float 0.000000e+00, float addrspace(1)* [[ARRAYIDX]], align 4
; CHECK-NEXT:    ret void
;
entry:
  %p = getelementptr inbounds %struct.S, %struct.S addrspace(1)* %Arg, i64 0, i32 0
  %p1 = load float*, float* addrspace(1)* %p, align 8
  %id = tail call i32 @llvm.amdgcn.workitem.id.x()
  %arrayidx = getelementptr inbounds float, float* %p1, i32 %id
  store float 0.000000e+00, float* %arrayidx, align 4
  ret void
}

@LDS = internal unnamed_addr addrspace(3) global [4 x float] undef, align 16

; GCN-LABEL: flat_ptr_arg:
; GCN-COUNT-2: global_load_dwordx2
; GCN:         global_load_dwordx4
; GCN:         global_store_dword
define amdgpu_kernel void @flat_ptr_arg(float** nocapture readonly noalias %Arg, float** nocapture noalias %Out, i32 %X) {
; CHECK-LABEL: @flat_ptr_arg(
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[OUT_GLOBAL:%.*]] = addrspacecast float** [[OUT:%.*]] to float* addrspace(1)*
; CHECK-NEXT:    [[ARG_GLOBAL:%.*]] = addrspacecast float** [[ARG:%.*]] to float* addrspace(1)*
; CHECK-NEXT:    [[I:%.*]] = tail call i32 @llvm.amdgcn.workitem.id.x()
; CHECK-NEXT:    [[IDXPROM:%.*]] = zext i32 [[I]] to i64
; CHECK-NEXT:    [[ARRAYIDX10:%.*]] = getelementptr inbounds float*, float* addrspace(1)* [[ARG_GLOBAL]], i64 [[IDXPROM]]
; CHECK-NEXT:    [[I1:%.*]] = load float*, float* addrspace(1)* [[ARRAYIDX10]], align 8
; CHECK-NEXT:    [[I1_GLOBAL:%.*]] = addrspacecast float* [[I1]] to float addrspace(1)*
; CHECK-NEXT:    [[I2:%.*]] = load float, float addrspace(1)* [[I1_GLOBAL]], align 4
; CHECK-NEXT:    [[ARRAYIDX512:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[X:%.*]]
; CHECK-NEXT:    store float [[I2]], float addrspace(3)* [[ARRAYIDX512]], align 4
; CHECK-NEXT:    [[ARRAYIDX3_1:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I1_GLOBAL]], i64 1
; CHECK-NEXT:    [[I3:%.*]] = load float, float addrspace(1)* [[ARRAYIDX3_1]], align 4
; CHECK-NEXT:    [[ADD_1:%.*]] = add nsw i32 [[X]], 1
; CHECK-NEXT:    [[ARRAYIDX512_1:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[ADD_1]]
; CHECK-NEXT:    store float [[I3]], float addrspace(3)* [[ARRAYIDX512_1]], align 4
; CHECK-NEXT:    [[ARRAYIDX3_2:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I1_GLOBAL]], i64 2
; CHECK-NEXT:    [[I4:%.*]] = load float, float addrspace(1)* [[ARRAYIDX3_2]], align 4
; CHECK-NEXT:    [[ADD_2:%.*]] = add nsw i32 [[X]], 2
; CHECK-NEXT:    [[ARRAYIDX512_2:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[ADD_2]]
; CHECK-NEXT:    store float [[I4]], float addrspace(3)* [[ARRAYIDX512_2]], align 4
; CHECK-NEXT:    [[ARRAYIDX3_3:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I1_GLOBAL]], i64 3
; CHECK-NEXT:    [[I5:%.*]] = load float, float addrspace(1)* [[ARRAYIDX3_3]], align 4
; CHECK-NEXT:    [[ADD_3:%.*]] = add nsw i32 [[X]], 3
; CHECK-NEXT:    [[ARRAYIDX512_3:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[ADD_3]]
; CHECK-NEXT:    store float [[I5]], float addrspace(3)* [[ARRAYIDX512_3]], align 4
; CHECK-NEXT:    [[SUB:%.*]] = add nsw i32 [[X]], -1
; CHECK-NEXT:    [[ARRAYIDX711:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[SUB]]
; CHECK-NEXT:    [[I6:%.*]] = load float, float addrspace(3)* [[ARRAYIDX711]], align 4
; CHECK-NEXT:    [[ARRAYIDX11:%.*]] = getelementptr inbounds float*, float* addrspace(1)* [[OUT_GLOBAL]], i64 [[IDXPROM]]
; CHECK-NEXT:    [[I7:%.*]] = load float*, float* addrspace(1)* [[ARRAYIDX11]], align 8
; CHECK-NEXT:    [[I7_GLOBAL:%.*]] = addrspacecast float* [[I7]] to float addrspace(1)*
; CHECK-NEXT:    [[IDXPROM8:%.*]] = sext i32 [[X]] to i64
; CHECK-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I7_GLOBAL]], i64 [[IDXPROM8]]
; CHECK-NEXT:    store float [[I6]], float addrspace(1)* [[ARRAYIDX9]], align 4
; CHECK-NEXT:    ret void
;
entry:
  %i = tail call i32 @llvm.amdgcn.workitem.id.x()
  %idxprom = zext i32 %i to i64
  %arrayidx10 = getelementptr inbounds float*, float** %Arg, i64 %idxprom
  %i1 = load float*, float** %arrayidx10, align 8
  %i2 = load float, float* %i1, align 4
  %arrayidx512 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %X
  store float %i2, float addrspace(3)* %arrayidx512, align 4
  %arrayidx3.1 = getelementptr inbounds float, float* %i1, i64 1
  %i3 = load float, float* %arrayidx3.1, align 4
  %add.1 = add nsw i32 %X, 1
  %arrayidx512.1 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %add.1
  store float %i3, float addrspace(3)* %arrayidx512.1, align 4
  %arrayidx3.2 = getelementptr inbounds float, float* %i1, i64 2
  %i4 = load float, float* %arrayidx3.2, align 4
  %add.2 = add nsw i32 %X, 2
  %arrayidx512.2 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %add.2
  store float %i4, float addrspace(3)* %arrayidx512.2, align 4
  %arrayidx3.3 = getelementptr inbounds float, float* %i1, i64 3
  %i5 = load float, float* %arrayidx3.3, align 4
  %add.3 = add nsw i32 %X, 3
  %arrayidx512.3 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %add.3
  store float %i5, float addrspace(3)* %arrayidx512.3, align 4
  %sub = add nsw i32 %X, -1
  %arrayidx711 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %sub
  %i6 = load float, float addrspace(3)* %arrayidx711, align 4
  %arrayidx11 = getelementptr inbounds float*, float** %Out, i64 %idxprom
  %i7 = load float*, float** %arrayidx11, align 8
  %idxprom8 = sext i32 %X to i64
  %arrayidx9 = getelementptr inbounds float, float* %i7, i64 %idxprom8
  store float %i6, float* %arrayidx9, align 4
  ret void
}

; GCN-LABEL: global_ptr_arg:
; GCN: global_load_dwordx2
; GCN: global_load_dwordx4
; GCN: global_store_dword
define amdgpu_kernel void @global_ptr_arg(float* addrspace(1)* nocapture readonly %Arg, i32 %X) {
; CHECK-LABEL: @global_ptr_arg(
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[I:%.*]] = tail call i32 @llvm.amdgcn.workitem.id.x()
; CHECK-NEXT:    [[IDXPROM:%.*]] = zext i32 [[I]] to i64
; CHECK-NEXT:    [[ARRAYIDX10:%.*]] = getelementptr inbounds float*, float* addrspace(1)* [[ARG:%.*]], i64 [[IDXPROM]]
; CHECK-NEXT:    [[I1:%.*]] = load float*, float* addrspace(1)* [[ARRAYIDX10]], align 8
; CHECK-NEXT:    [[I1_GLOBAL:%.*]] = addrspacecast float* [[I1]] to float addrspace(1)*
; CHECK-NEXT:    [[I2:%.*]] = load float, float addrspace(1)* [[I1_GLOBAL]], align 4
; CHECK-NEXT:    [[ARRAYIDX512:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[X:%.*]]
; CHECK-NEXT:    store float [[I2]], float addrspace(3)* [[ARRAYIDX512]], align 4
; CHECK-NEXT:    [[ARRAYIDX3_1:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I1_GLOBAL]], i64 1
; CHECK-NEXT:    [[I3:%.*]] = load float, float addrspace(1)* [[ARRAYIDX3_1]], align 4
; CHECK-NEXT:    [[ADD_1:%.*]] = add nsw i32 [[X]], 1
; CHECK-NEXT:    [[ARRAYIDX512_1:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[ADD_1]]
; CHECK-NEXT:    store float [[I3]], float addrspace(3)* [[ARRAYIDX512_1]], align 4
; CHECK-NEXT:    [[ARRAYIDX3_2:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I1_GLOBAL]], i64 2
; CHECK-NEXT:    [[I4:%.*]] = load float, float addrspace(1)* [[ARRAYIDX3_2]], align 4
; CHECK-NEXT:    [[ADD_2:%.*]] = add nsw i32 [[X]], 2
; CHECK-NEXT:    [[ARRAYIDX512_2:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[ADD_2]]
; CHECK-NEXT:    store float [[I4]], float addrspace(3)* [[ARRAYIDX512_2]], align 4
; CHECK-NEXT:    [[ARRAYIDX3_3:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I1_GLOBAL]], i64 3
; CHECK-NEXT:    [[I5:%.*]] = load float, float addrspace(1)* [[ARRAYIDX3_3]], align 4
; CHECK-NEXT:    [[ADD_3:%.*]] = add nsw i32 [[X]], 3
; CHECK-NEXT:    [[ARRAYIDX512_3:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[ADD_3]]
; CHECK-NEXT:    store float [[I5]], float addrspace(3)* [[ARRAYIDX512_3]], align 4
; CHECK-NEXT:    [[SUB:%.*]] = add nsw i32 [[X]], -1
; CHECK-NEXT:    [[ARRAYIDX711:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[SUB]]
; CHECK-NEXT:    [[I6:%.*]] = load float, float addrspace(3)* [[ARRAYIDX711]], align 4
; CHECK-NEXT:    [[IDXPROM8:%.*]] = sext i32 [[X]] to i64
; CHECK-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I1_GLOBAL]], i64 [[IDXPROM8]]
; CHECK-NEXT:    store float [[I6]], float addrspace(1)* [[ARRAYIDX9]], align 4
; CHECK-NEXT:    ret void
;
entry:
  %i = tail call i32 @llvm.amdgcn.workitem.id.x()
  %idxprom = zext i32 %i to i64
  %arrayidx10 = getelementptr inbounds float*, float* addrspace(1)* %Arg, i64 %idxprom
  %i1 = load float*, float* addrspace(1)* %arrayidx10, align 8
  %i2 = load float, float* %i1, align 4
  %arrayidx512 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %X
  store float %i2, float addrspace(3)* %arrayidx512, align 4
  %arrayidx3.1 = getelementptr inbounds float, float* %i1, i64 1
  %i3 = load float, float* %arrayidx3.1, align 4
  %add.1 = add nsw i32 %X, 1
  %arrayidx512.1 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %add.1
  store float %i3, float addrspace(3)* %arrayidx512.1, align 4
  %arrayidx3.2 = getelementptr inbounds float, float* %i1, i64 2
  %i4 = load float, float* %arrayidx3.2, align 4
  %add.2 = add nsw i32 %X, 2
  %arrayidx512.2 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %add.2
  store float %i4, float addrspace(3)* %arrayidx512.2, align 4
  %arrayidx3.3 = getelementptr inbounds float, float* %i1, i64 3
  %i5 = load float, float* %arrayidx3.3, align 4
  %add.3 = add nsw i32 %X, 3
  %arrayidx512.3 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %add.3
  store float %i5, float addrspace(3)* %arrayidx512.3, align 4
  %sub = add nsw i32 %X, -1
  %arrayidx711 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %sub
  %i6 = load float, float addrspace(3)* %arrayidx711, align 4
  %idxprom8 = sext i32 %X to i64
  %arrayidx9 = getelementptr inbounds float, float* %i1, i64 %idxprom8
  store float %i6, float* %arrayidx9, align 4
  ret void
}

; GCN-LABEL: global_ptr_arg_clobbered:
; GCN: global_store_dwordx2
; GCN: global_load_dwordx2
; GCN: flat_load_dword
; GCN: flat_store_dword
define amdgpu_kernel void @global_ptr_arg_clobbered(float* addrspace(1)* nocapture readonly %Arg, i32 %X) {
; CHECK-LABEL: @global_ptr_arg_clobbered(
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[I:%.*]] = tail call i32 @llvm.amdgcn.workitem.id.x()
; CHECK-NEXT:    [[IDXPROM:%.*]] = zext i32 [[I]] to i64
; CHECK-NEXT:    [[ARRAYIDX10:%.*]] = getelementptr inbounds float*, float* addrspace(1)* [[ARG:%.*]], i64 [[IDXPROM]]
; CHECK-NEXT:    [[ARRAYIDX11:%.*]] = getelementptr inbounds float*, float* addrspace(1)* [[ARRAYIDX10]], i32 [[X:%.*]]
; CHECK-NEXT:    store float* null, float* addrspace(1)* [[ARRAYIDX11]], align 4
; CHECK-NEXT:    [[I1:%.*]] = load float*, float* addrspace(1)* [[ARRAYIDX10]], align 8
; CHECK-NEXT:    [[I2:%.*]] = load float, float* [[I1]], align 4
; CHECK-NEXT:    [[ARRAYIDX512:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[X]]
; CHECK-NEXT:    store float [[I2]], float addrspace(3)* [[ARRAYIDX512]], align 4
; CHECK-NEXT:    [[SUB:%.*]] = add nsw i32 [[X]], -1
; CHECK-NEXT:    [[ARRAYIDX711:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[SUB]]
; CHECK-NEXT:    [[I6:%.*]] = load float, float addrspace(3)* [[ARRAYIDX711]], align 4
; CHECK-NEXT:    [[IDXPROM8:%.*]] = sext i32 [[X]] to i64
; CHECK-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds float, float* [[I1]], i64 [[IDXPROM8]]
; CHECK-NEXT:    store float [[I6]], float* [[ARRAYIDX9]], align 4
; CHECK-NEXT:    ret void
;
entry:
  %i = tail call i32 @llvm.amdgcn.workitem.id.x()
  %idxprom = zext i32 %i to i64
  %arrayidx10 = getelementptr inbounds float*, float* addrspace(1)* %Arg, i64 %idxprom
  %arrayidx11 = getelementptr inbounds float*, float* addrspace(1)* %arrayidx10, i32 %X
  store float* null, float* addrspace(1)* %arrayidx11, align 4
  %i1 = load float*, float* addrspace(1)* %arrayidx10, align 8
  %i2 = load float, float* %i1, align 4
  %arrayidx512 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %X
  store float %i2, float addrspace(3)* %arrayidx512, align 4
  %sub = add nsw i32 %X, -1
  %arrayidx711 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %sub
  %i6 = load float, float addrspace(3)* %arrayidx711, align 4
  %idxprom8 = sext i32 %X to i64
  %arrayidx9 = getelementptr inbounds float, float* %i1, i64 %idxprom8
  store float %i6, float* %arrayidx9, align 4
  ret void
}

; GCN-LABEL: global_ptr_arg_clobbered_after_load:
; GCN: global_load_dwordx2
; GCN: global_store_dwordx2
; GCN: global_load_dword
; GCN: global_store_dword
define amdgpu_kernel void @global_ptr_arg_clobbered_after_load(float* addrspace(1)* nocapture readonly %Arg, i32 %X) {
; CHECK-LABEL: @global_ptr_arg_clobbered_after_load(
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[I:%.*]] = tail call i32 @llvm.amdgcn.workitem.id.x()
; CHECK-NEXT:    [[IDXPROM:%.*]] = zext i32 [[I]] to i64
; CHECK-NEXT:    [[ARRAYIDX10:%.*]] = getelementptr inbounds float*, float* addrspace(1)* [[ARG:%.*]], i64 [[IDXPROM]]
; CHECK-NEXT:    [[I1:%.*]] = load float*, float* addrspace(1)* [[ARRAYIDX10]], align 8
; CHECK-NEXT:    [[I1_GLOBAL:%.*]] = addrspacecast float* [[I1]] to float addrspace(1)*
; CHECK-NEXT:    [[ARRAYIDX11:%.*]] = getelementptr inbounds float*, float* addrspace(1)* [[ARRAYIDX10]], i32 [[X:%.*]]
; CHECK-NEXT:    store float* null, float* addrspace(1)* [[ARRAYIDX11]], align 4
; CHECK-NEXT:    [[I2:%.*]] = load float, float addrspace(1)* [[I1_GLOBAL]], align 4
; CHECK-NEXT:    [[ARRAYIDX512:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[X]]
; CHECK-NEXT:    store float [[I2]], float addrspace(3)* [[ARRAYIDX512]], align 4
; CHECK-NEXT:    [[SUB:%.*]] = add nsw i32 [[X]], -1
; CHECK-NEXT:    [[ARRAYIDX711:%.*]] = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 [[SUB]]
; CHECK-NEXT:    [[I6:%.*]] = load float, float addrspace(3)* [[ARRAYIDX711]], align 4
; CHECK-NEXT:    [[IDXPROM8:%.*]] = sext i32 [[X]] to i64
; CHECK-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds float, float addrspace(1)* [[I1_GLOBAL]], i64 [[IDXPROM8]]
; CHECK-NEXT:    store float [[I6]], float addrspace(1)* [[ARRAYIDX9]], align 4
; CHECK-NEXT:    ret void
;
entry:
  %i = tail call i32 @llvm.amdgcn.workitem.id.x()
  %idxprom = zext i32 %i to i64
  %arrayidx10 = getelementptr inbounds float*, float* addrspace(1)* %Arg, i64 %idxprom
  %i1 = load float*, float* addrspace(1)* %arrayidx10, align 8
  %arrayidx11 = getelementptr inbounds float*, float* addrspace(1)* %arrayidx10, i32 %X
  store float* null, float* addrspace(1)* %arrayidx11, align 4
  %i2 = load float, float* %i1, align 4
  %arrayidx512 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %X
  store float %i2, float addrspace(3)* %arrayidx512, align 4
  %sub = add nsw i32 %X, -1
  %arrayidx711 = getelementptr inbounds [4 x float], [4 x float] addrspace(3)* @LDS, i32 0, i32 %sub
  %i6 = load float, float addrspace(3)* %arrayidx711, align 4
  %idxprom8 = sext i32 %X to i64
  %arrayidx9 = getelementptr inbounds float, float* %i1, i64 %idxprom8
  store float %i6, float* %arrayidx9, align 4
  ret void
}

declare i32 @llvm.amdgcn.workitem.id.x()