1;RUN: llc < %s -march=amdgcn -mcpu=verde -verify-machineinstrs | FileCheck -check-prefixes=GCN,VERDE %s
2;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck -check-prefix=GCN %s
3
4; GCN-LABEL: {{^}}tbuffer_store:
5; GCN: tbuffer_store_format_xyzw v[0:3], off, s[0:3], dfmt:12, nfmt:2, 0
6; GCN: tbuffer_store_format_xyzw v[4:7], off, s[0:3], dfmt:13, nfmt:3, 0 glc
7; GCN: tbuffer_store_format_xyzw v[8:11], off, s[0:3], dfmt:14, nfmt:4, 0 slc
8; GCN: tbuffer_store_format_xyzw v[8:11], off, s[0:3], dfmt:14, nfmt:4, 0
9define amdgpu_ps void @tbuffer_store(<4 x i32> inreg, <4 x float>, <4 x float>, <4 x float>) {
10main_body:
11  %in1 = bitcast <4 x float> %1 to <4 x i32>
12  %in2 = bitcast <4 x float> %2 to <4 x i32>
13  %in3 = bitcast <4 x float> %3 to <4 x i32>
14  call void @llvm.amdgcn.raw.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 0, i32 44, i32 0)
15  call void @llvm.amdgcn.raw.tbuffer.store.v4i32(<4 x i32> %in2, <4 x i32> %0, i32 0, i32 0, i32 61, i32 1)
16  call void @llvm.amdgcn.raw.tbuffer.store.v4i32(<4 x i32> %in3, <4 x i32> %0, i32 0, i32 0, i32 78, i32 2)
17  call void @llvm.amdgcn.raw.tbuffer.store.v4f32(<4 x float> %3, <4 x i32> %0, i32 0, i32 0, i32 78, i32 0)
18  ret void
19}
20
21; GCN-LABEL: {{^}}tbuffer_store_immoffs:
22; GCN: tbuffer_store_format_xyzw v[0:3], off, s[0:3], dfmt:5, nfmt:7, 0 offset:42
23define amdgpu_ps void @tbuffer_store_immoffs(<4 x i32> inreg, <4 x float>) {
24main_body:
25  %in1 = bitcast <4 x float> %1 to <4 x i32>
26  call void @llvm.amdgcn.raw.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 42, i32 0, i32 117, i32 0)
27  ret void
28}
29
30; GCN-LABEL: {{^}}tbuffer_store_scalar_and_imm_offs:
31; GCN: tbuffer_store_format_xyzw v[0:3], off, s[0:3], dfmt:5, nfmt:7, {{s[0-9]+}} offset:42
32define amdgpu_ps void @tbuffer_store_scalar_and_imm_offs(<4 x i32> inreg, <4 x float> %vdata, i32 inreg %soffset) {
33main_body:
34  %in1 = bitcast <4 x float> %vdata to <4 x i32>
35  call void @llvm.amdgcn.raw.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 42, i32 %soffset, i32 117, i32 0)
36  ret void
37}
38
39; GCN-LABEL: {{^}}buffer_store_ofs:
40; GCN: tbuffer_store_format_xyzw v[0:3], v4, s[0:3], dfmt:3, nfmt:7, 0 offen
41define amdgpu_ps void @buffer_store_ofs(<4 x i32> inreg, <4 x float> %vdata, i32 %voffset) {
42main_body:
43  %in1 = bitcast <4 x float> %vdata to <4 x i32>
44  call void @llvm.amdgcn.raw.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 %voffset, i32 0, i32 115, i32 0)
45  ret void
46}
47
48; GCN-LABEL: {{^}}buffer_store_x1:
49; GCN: tbuffer_store_format_x v0, off, s[0:3], dfmt:13, nfmt:7, 0
50define amdgpu_ps void @buffer_store_x1(<4 x i32> inreg %rsrc, float %data) {
51main_body:
52  %data.i = bitcast float %data to i32
53  call void @llvm.amdgcn.raw.tbuffer.store.i32(i32 %data.i, <4 x i32> %rsrc, i32 0, i32 0, i32 125, i32 0)
54  ret void
55}
56
57; GCN-LABEL: {{^}}buffer_store_x2:
58; GCN: tbuffer_store_format_xy v[0:1], off, s[0:3], dfmt:1, nfmt:2, 0
59define amdgpu_ps void @buffer_store_x2(<4 x i32> inreg %rsrc, <2 x float> %data) {
60main_body:
61  %data.i = bitcast <2 x float> %data to <2 x i32>
62  call void @llvm.amdgcn.raw.tbuffer.store.v2i32(<2 x i32> %data.i, <4 x i32> %rsrc, i32 0, i32 0, i32 33, i32 0)
63  ret void
64}
65
66declare void @llvm.amdgcn.raw.tbuffer.store.i32(i32, <4 x i32>, i32, i32, i32, i32) #0
67declare void @llvm.amdgcn.raw.tbuffer.store.v2i32(<2 x i32>, <4 x i32>, i32, i32, i32, i32) #0
68declare void @llvm.amdgcn.raw.tbuffer.store.v4i32(<4 x i32>, <4 x i32>, i32, i32, i32, i32) #0
69declare void @llvm.amdgcn.raw.tbuffer.store.v4f32(<4 x float>, <4 x i32>, i32, i32, i32, i32) #0
70declare <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32>, i32, i1, i1) #1
71
72attributes #0 = { nounwind }
73attributes #1 = { nounwind readonly }
74
75
76