|
121
|
1 ; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-NOHSA -check-prefix=FUNC %s
|
|
|
2 ; RUN: llc -amdgpu-scalarize-global-loads=false -mtriple=amdgcn--amdhsa -mcpu=kaveri -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-HSA -check-prefix=FUNC %s
|
|
|
3 ; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GCN-NOHSA -check-prefix=FUNC %s
|
|
120
|
4
|
|
121
|
5 ; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
|
|
|
6 ; RUN: llc -amdgpu-scalarize-global-loads=false -march=r600 -mcpu=cayman < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
|
|
120
|
7
|
|
|
8 ; FUNC-LABEL: {{^}}global_load_i64:
|
|
|
9 ; GCN-NOHSA: buffer_load_dwordx2 [[VAL:v\[[0-9]+:[0-9]+\]]]
|
|
|
10 ; GCN-NOHSA: buffer_store_dwordx2 [[VAL]]
|
|
|
11
|
|
|
12 ; GCN-HSA: flat_load_dwordx2 [[VAL:v\[[0-9]+:[0-9]+\]]]
|
|
|
13 ; GCN-HSA: flat_store_dwordx2 {{v\[[0-9]+:[0-9]+\]}}, [[VAL]]
|
|
|
14
|
|
|
15 ; EG: VTX_READ_64
|
|
121
|
16 define amdgpu_kernel void @global_load_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %in) #0 {
|
|
120
|
17 %ld = load i64, i64 addrspace(1)* %in
|
|
|
18 store i64 %ld, i64 addrspace(1)* %out
|
|
|
19 ret void
|
|
|
20 }
|
|
|
21
|
|
|
22 ; FUNC-LABEL: {{^}}global_load_v2i64:
|
|
|
23 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
24 ; GCN-HSA: flat_load_dwordx4
|
|
|
25
|
|
|
26 ; EG: VTX_READ_128
|
|
121
|
27 define amdgpu_kernel void @global_load_v2i64(<2 x i64> addrspace(1)* %out, <2 x i64> addrspace(1)* %in) #0 {
|
|
120
|
28 entry:
|
|
|
29 %ld = load <2 x i64>, <2 x i64> addrspace(1)* %in
|
|
|
30 store <2 x i64> %ld, <2 x i64> addrspace(1)* %out
|
|
|
31 ret void
|
|
|
32 }
|
|
|
33
|
|
|
34 ; FUNC-LABEL: {{^}}global_load_v3i64:
|
|
|
35 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
36 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
37
|
|
|
38 ; GCN-HSA: flat_load_dwordx4
|
|
|
39 ; GCN-HSA: flat_load_dwordx4
|
|
|
40
|
|
|
41 ; EG: VTX_READ_128
|
|
|
42 ; EG: VTX_READ_128
|
|
121
|
43 define amdgpu_kernel void @global_load_v3i64(<3 x i64> addrspace(1)* %out, <3 x i64> addrspace(1)* %in) #0 {
|
|
120
|
44 entry:
|
|
|
45 %ld = load <3 x i64>, <3 x i64> addrspace(1)* %in
|
|
|
46 store <3 x i64> %ld, <3 x i64> addrspace(1)* %out
|
|
|
47 ret void
|
|
|
48 }
|
|
|
49
|
|
|
50 ; FUNC-LABEL: {{^}}global_load_v4i64:
|
|
|
51 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
52 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
53
|
|
|
54 ; GCN-HSA: flat_load_dwordx4
|
|
|
55 ; GCN-HSA: flat_load_dwordx4
|
|
|
56
|
|
|
57 ; EG: VTX_READ_128
|
|
|
58 ; EG: VTX_READ_128
|
|
121
|
59 define amdgpu_kernel void @global_load_v4i64(<4 x i64> addrspace(1)* %out, <4 x i64> addrspace(1)* %in) #0 {
|
|
120
|
60 entry:
|
|
|
61 %ld = load <4 x i64>, <4 x i64> addrspace(1)* %in
|
|
|
62 store <4 x i64> %ld, <4 x i64> addrspace(1)* %out
|
|
|
63 ret void
|
|
|
64 }
|
|
|
65
|
|
|
66 ; FUNC-LABEL: {{^}}global_load_v8i64:
|
|
|
67 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
68 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
69 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
70 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
71
|
|
|
72 ; GCN-HSA: flat_load_dwordx4
|
|
|
73 ; GCN-HSA: flat_load_dwordx4
|
|
|
74 ; GCN-HSA: flat_load_dwordx4
|
|
|
75 ; GCN-HSA: flat_load_dwordx4
|
|
|
76
|
|
|
77 ; EG: VTX_READ_128
|
|
|
78 ; EG: VTX_READ_128
|
|
|
79 ; EG: VTX_READ_128
|
|
|
80 ; EG: VTX_READ_128
|
|
121
|
81 define amdgpu_kernel void @global_load_v8i64(<8 x i64> addrspace(1)* %out, <8 x i64> addrspace(1)* %in) #0 {
|
|
120
|
82 entry:
|
|
|
83 %ld = load <8 x i64>, <8 x i64> addrspace(1)* %in
|
|
|
84 store <8 x i64> %ld, <8 x i64> addrspace(1)* %out
|
|
|
85 ret void
|
|
|
86 }
|
|
|
87
|
|
|
88 ; FUNC-LABEL: {{^}}global_load_v16i64:
|
|
|
89 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
90 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
91 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
92 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
93 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
94 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
95 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
96 ; GCN-NOHSA: buffer_load_dwordx4
|
|
|
97
|
|
|
98 ; GCN-HSA: flat_load_dwordx4
|
|
|
99 ; GCN-HSA: flat_load_dwordx4
|
|
|
100 ; GCN-HSA: flat_load_dwordx4
|
|
|
101 ; GCN-HSA: flat_load_dwordx4
|
|
|
102 ; GCN-HSA: flat_load_dwordx4
|
|
|
103 ; GCN-HSA: flat_load_dwordx4
|
|
|
104 ; GCN-HSA: flat_load_dwordx4
|
|
|
105 ; GCN-HSA: flat_load_dwordx4
|
|
|
106
|
|
|
107 ; EG: VTX_READ_128
|
|
|
108 ; EG: VTX_READ_128
|
|
|
109 ; EG: VTX_READ_128
|
|
|
110 ; EG: VTX_READ_128
|
|
|
111 ; EG: VTX_READ_128
|
|
|
112 ; EG: VTX_READ_128
|
|
|
113 ; EG: VTX_READ_128
|
|
|
114 ; EG: VTX_READ_128
|
|
121
|
115 define amdgpu_kernel void @global_load_v16i64(<16 x i64> addrspace(1)* %out, <16 x i64> addrspace(1)* %in) #0 {
|
|
120
|
116 entry:
|
|
|
117 %ld = load <16 x i64>, <16 x i64> addrspace(1)* %in
|
|
|
118 store <16 x i64> %ld, <16 x i64> addrspace(1)* %out
|
|
|
119 ret void
|
|
|
120 }
|
|
|
121
|
|
|
122 attributes #0 = { nounwind }
|
