File: cp-async-bulk-tensor-s2g.ll

package info (click to toggle)
llvm-toolchain-20 1%3A20.1.8-1
  • links: PTS, VCS
  • area: main
  • in suites: experimental
  • size: 2,111,696 kB
  • sloc: cpp: 7,438,781; ansic: 1,393,871; asm: 1,012,926; python: 241,771; f90: 86,635; objc: 75,411; lisp: 42,144; pascal: 17,286; sh: 8,596; ml: 5,082; perl: 4,730; makefile: 3,591; awk: 3,523; javascript: 2,251; xml: 892; fortran: 672
file content (228 lines) | stat: -rw-r--r-- 16,930 bytes parent folder | download | duplicates (2) 
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
 
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_90 -mattr=+ptx80| FileCheck --check-prefixes=CHECK-PTX64 %s
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_90 -mattr=+ptx80 --nvptx-short-ptr| FileCheck --check-prefixes=CHECK-PTX-SHARED32 %s
; RUN: %if ptxas-12.3 %{ llc < %s -mtriple=nvptx64 -mcpu=sm_90 -mattr=+ptx80| %ptxas-verify -arch=sm_90 %}
; RUN: %if ptxas-12.3 %{ llc < %s -mtriple=nvptx64 -mcpu=sm_90 -mattr=+ptx80 --nvptx-short-ptr| %ptxas-verify -arch=sm_90 %}

target triple = "nvptx64-nvidia-cuda"

declare void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.1d(ptr addrspace(3) %s, ptr %tm, i32 %d0, i64 %ch, i1 %flag);
declare void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.2d(ptr addrspace(3) %s, ptr %tm, i32 %d0, i32 %d1, i64 %ch, i1 %flag);
declare void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.3d(ptr addrspace(3) %s, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i64 %ch, i1 %flag);
declare void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.4d(ptr addrspace(3) %s, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i64 %ch, i1 %flag);
declare void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.5d(ptr addrspace(3) %s, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i64 %ch, i1 %flag);

declare void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.3d(ptr addrspace(3) %s, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i64 %ch, i1 %flag);
declare void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.4d(ptr addrspace(3) %s, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i64 %ch, i1 %flag);
declare void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.5d(ptr addrspace(3) %s, ptr %tm, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i64 %ch, i1 %flag);

; CHECK-LABEL: cp_async_bulk_tensor_s2g_tile_1d
define void @cp_async_bulk_tensor_s2g_tile_1d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_s2g_tile_1d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b32 %r<2>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<4>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_tile_1d_param_0];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_tile_1d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_tile_1d_param_2];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.1d.global.shared::cta.tile.bulk_group [%rd2, {%r1}], [%rd1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_s2g_tile_1d_param_3];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.1d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd2, {%r1}], [%rd1], %rd3;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_s2g_tile_1d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<3>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_tile_1d_param_0];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_tile_1d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_tile_1d_param_2];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.1d.global.shared::cta.tile.bulk_group [%rd1, {%r2}], [%r1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_tile_1d_param_3];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.1d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd1, {%r2}], [%r1], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.1d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i64 undef, i1 0)
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.1d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i64 %ch, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_s2g_tile_2d
define void @cp_async_bulk_tensor_s2g_tile_2d(i32 %flag, ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_s2g_tile_2d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b32 %r<3>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<4>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_tile_2d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_tile_2d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_tile_2d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_tile_2d_param_4];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.2d.global.shared::cta.tile.bulk_group [%rd2, {%r1, %r2}], [%rd1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_s2g_tile_2d_param_5];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.2d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd2, {%r1, %r2}], [%rd1], %rd3;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_s2g_tile_2d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<4>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_tile_2d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_tile_2d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_tile_2d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_s2g_tile_2d_param_4];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.2d.global.shared::cta.tile.bulk_group [%rd1, {%r2, %r3}], [%r1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_tile_2d_param_5];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.2d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd1, {%r2, %r3}], [%r1], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.2d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i64 undef, i1 0)
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.2d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i64 %ch, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_s2g_3d
define void @cp_async_bulk_tensor_s2g_3d(i32 %flag, ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_s2g_3d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b32 %r<4>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<4>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_3d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_3d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_3d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_3d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_s2g_3d_param_5];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.global.shared::cta.tile.bulk_group [%rd2, {%r1, %r2, %r3}], [%rd1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_s2g_3d_param_6];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd2, {%r1, %r2, %r3}], [%rd1], %rd3;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.global.shared::cta.im2col_no_offs.bulk_group [%rd2, {%r1, %r2, %r3}], [%rd1];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.3d.global.shared::cta.im2col_no_offs.bulk_group.L2::cache_hint [%rd2, {%r1, %r2, %r3}], [%rd1], %rd3;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_s2g_3d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<5>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_3d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_3d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_3d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_s2g_3d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_s2g_3d_param_5];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.global.shared::cta.tile.bulk_group [%rd1, {%r2, %r3, %r4}], [%r1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_3d_param_6];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd1, {%r2, %r3, %r4}], [%r1], %rd2;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.global.shared::cta.im2col_no_offs.bulk_group [%rd1, {%r2, %r3, %r4}], [%r1];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.3d.global.shared::cta.im2col_no_offs.bulk_group.L2::cache_hint [%rd1, {%r2, %r3, %r4}], [%r1], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.3d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i64 undef, i1 0)
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.3d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i64 %ch, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.3d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i64 undef, i1 0)
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.3d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i64 %ch, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_s2g_4d
define void @cp_async_bulk_tensor_s2g_4d(i32 %flag, ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_s2g_4d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b32 %r<5>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<4>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_4d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_4d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_4d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_4d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_s2g_4d_param_5];
; CHECK-PTX64-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_s2g_4d_param_6];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.global.shared::cta.tile.bulk_group [%rd2, {%r1, %r2, %r3, %r4}], [%rd1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_s2g_4d_param_7];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd2, {%r1, %r2, %r3, %r4}], [%rd1], %rd3;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.global.shared::cta.im2col_no_offs.bulk_group [%rd2, {%r1, %r2, %r3, %r4}], [%rd1];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.4d.global.shared::cta.im2col_no_offs.bulk_group.L2::cache_hint [%rd2, {%r1, %r2, %r3, %r4}], [%rd1], %rd3;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_s2g_4d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<6>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_4d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_4d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_4d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_s2g_4d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_s2g_4d_param_5];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_s2g_4d_param_6];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.global.shared::cta.tile.bulk_group [%rd1, {%r2, %r3, %r4, %r5}], [%r1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_4d_param_7];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd1, {%r2, %r3, %r4, %r5}], [%r1], %rd2;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.global.shared::cta.im2col_no_offs.bulk_group [%rd1, {%r2, %r3, %r4, %r5}], [%r1];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.4d.global.shared::cta.im2col_no_offs.bulk_group.L2::cache_hint [%rd1, {%r2, %r3, %r4, %r5}], [%r1], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.4d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i64 undef, i1 0)
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.4d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i64 %ch, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.4d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i64 undef, i1 0)
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.4d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i64 %ch, i1 1)
  ret void
}

; CHECK-LABEL: cp_async_bulk_tensor_s2g_5d
define void @cp_async_bulk_tensor_s2g_5d(i32 %flag, ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i64 %ch) {
; CHECK-PTX64-LABEL: cp_async_bulk_tensor_s2g_5d(
; CHECK-PTX64:       {
; CHECK-PTX64-NEXT:    .reg .b32 %r<6>;
; CHECK-PTX64-NEXT:    .reg .b64 %rd<4>;
; CHECK-PTX64-EMPTY:
; CHECK-PTX64-NEXT:  // %bb.0:
; CHECK-PTX64-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_5d_param_1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_5d_param_2];
; CHECK-PTX64-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_5d_param_3];
; CHECK-PTX64-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_5d_param_4];
; CHECK-PTX64-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_s2g_5d_param_5];
; CHECK-PTX64-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_s2g_5d_param_6];
; CHECK-PTX64-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_s2g_5d_param_7];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.global.shared::cta.tile.bulk_group [%rd2, {%r1, %r2, %r3, %r4, %r5}], [%rd1];
; CHECK-PTX64-NEXT:    ld.param.u64 %rd3, [cp_async_bulk_tensor_s2g_5d_param_8];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd2, {%r1, %r2, %r3, %r4, %r5}], [%rd1], %rd3;
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.global.shared::cta.im2col_no_offs.bulk_group [%rd2, {%r1, %r2, %r3, %r4, %r5}], [%rd1];
; CHECK-PTX64-NEXT:    cp.async.bulk.tensor.5d.global.shared::cta.im2col_no_offs.bulk_group.L2::cache_hint [%rd2, {%r1, %r2, %r3, %r4, %r5}], [%rd1], %rd3;
; CHECK-PTX64-NEXT:    ret;
;
; CHECK-PTX-SHARED32-LABEL: cp_async_bulk_tensor_s2g_5d(
; CHECK-PTX-SHARED32:       {
; CHECK-PTX-SHARED32-NEXT:    .reg .b32 %r<7>;
; CHECK-PTX-SHARED32-NEXT:    .reg .b64 %rd<3>;
; CHECK-PTX-SHARED32-EMPTY:
; CHECK-PTX-SHARED32-NEXT:  // %bb.0:
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r1, [cp_async_bulk_tensor_s2g_5d_param_1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd1, [cp_async_bulk_tensor_s2g_5d_param_2];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r2, [cp_async_bulk_tensor_s2g_5d_param_3];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r3, [cp_async_bulk_tensor_s2g_5d_param_4];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r4, [cp_async_bulk_tensor_s2g_5d_param_5];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r5, [cp_async_bulk_tensor_s2g_5d_param_6];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u32 %r6, [cp_async_bulk_tensor_s2g_5d_param_7];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.global.shared::cta.tile.bulk_group [%rd1, {%r2, %r3, %r4, %r5, %r6}], [%r1];
; CHECK-PTX-SHARED32-NEXT:    ld.param.u64 %rd2, [cp_async_bulk_tensor_s2g_5d_param_8];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.global.shared::cta.tile.bulk_group.L2::cache_hint [%rd1, {%r2, %r3, %r4, %r5, %r6}], [%r1], %rd2;
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.global.shared::cta.im2col_no_offs.bulk_group [%rd1, {%r2, %r3, %r4, %r5, %r6}], [%r1];
; CHECK-PTX-SHARED32-NEXT:    cp.async.bulk.tensor.5d.global.shared::cta.im2col_no_offs.bulk_group.L2::cache_hint [%rd1, {%r2, %r3, %r4, %r5, %r6}], [%r1], %rd2;
; CHECK-PTX-SHARED32-NEXT:    ret;
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.5d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i64 undef, i1 0)
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.tile.5d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i64 %ch, i1 1)

  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.5d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i64 undef, i1 0)
  tail call void @llvm.nvvm.cp.async.bulk.tensor.s2g.im2col.5d(ptr addrspace(3) %src, ptr %tmap, i32 %d0, i32 %d1, i32 %d2, i32 %d3, i32 %d4, i64 %ch, i1 1)
  ret void
}