150
|
1 ;RUN: llc < %s -march=amdgcn -mcpu=verde -verify-machineinstrs | FileCheck -check-prefixes=GCN,VERDE,PREGFX10 %s
|
|
2 ;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck -check-prefixes=GCN,PREGFX10 %s
|
|
3 ;RUN: llc < %s -march=amdgcn -mcpu=gfx1010 -verify-machineinstrs | FileCheck -check-prefixes=GCN,GFX10 %s
|
|
4
|
|
5 ; GCN-LABEL: {{^}}tbuffer_store:
|
|
6 ; GCN: v_mov_b32_e32 [[ZEROREG:v[0-9]+]], 0
|
|
7 ; PREGFX10: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], dfmt:12, nfmt:2, 0 idxen
|
|
8 ; PREGFX10: tbuffer_store_format_xyzw v[4:7], [[ZEROREG]], s[0:3], dfmt:13, nfmt:3, 0 idxen glc
|
|
9 ; PREGFX10: tbuffer_store_format_xyzw v[8:11], [[ZEROREG]], s[0:3], dfmt:14, nfmt:4, 0 idxen slc
|
|
10 ; PREGFX10: tbuffer_store_format_xyzw v[8:11], [[ZEROREG]], s[0:3], dfmt:14, nfmt:4, 0 idxen glc
|
|
11 ; GFX10: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], format:44, 0 idxen
|
|
12 ; GFX10: tbuffer_store_format_xyzw v[4:7], [[ZEROREG]], s[0:3], format:61, 0 idxen glc
|
|
13 ; GFX10: tbuffer_store_format_xyzw v[8:11], [[ZEROREG]], s[0:3], format:78, 0 idxen slc
|
|
14 ; GFX10: tbuffer_store_format_xyzw v[8:11], [[ZEROREG]], s[0:3], format:78, 0 idxen glc dlc
|
|
15 define amdgpu_ps void @tbuffer_store(<4 x i32> inreg, <4 x float>, <4 x float>, <4 x float>) {
|
|
16 main_body:
|
|
17 %in1 = bitcast <4 x float> %1 to <4 x i32>
|
|
18 %in2 = bitcast <4 x float> %2 to <4 x i32>
|
|
19 %in3 = bitcast <4 x float> %3 to <4 x i32>
|
|
20 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 0, i32 0, i32 44, i32 0)
|
|
21 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in2, <4 x i32> %0, i32 0, i32 0, i32 0, i32 61, i32 1)
|
|
22 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in3, <4 x i32> %0, i32 0, i32 0, i32 0, i32 78, i32 2)
|
|
23 call void @llvm.amdgcn.struct.tbuffer.store.v4f32(<4 x float> %3, <4 x i32> %0, i32 0, i32 0, i32 0, i32 78, i32 5)
|
|
24 ret void
|
|
25 }
|
|
26
|
|
27 ; GCN-LABEL: {{^}}tbuffer_store_immoffs:
|
|
28 ; GCN: v_mov_b32_e32 [[ZEROREG:v[0-9]+]], 0
|
|
29 ; PREGFX10: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], dfmt:5, nfmt:7, 0 idxen offset:42
|
|
30 ; GFX10: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], format:117, 0 idxen offset:42
|
|
31 define amdgpu_ps void @tbuffer_store_immoffs(<4 x i32> inreg, <4 x float>) {
|
|
32 main_body:
|
|
33 %in1 = bitcast <4 x float> %1 to <4 x i32>
|
|
34 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 42, i32 0, i32 117, i32 0)
|
|
35 ret void
|
|
36 }
|
|
37
|
|
38 ; GCN-LABEL: {{^}}tbuffer_store_scalar_and_imm_offs:
|
|
39 ; GCN: v_mov_b32_e32 [[ZEROREG:v[0-9]+]], 0
|
|
40 ; PREGFX10: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], dfmt:5, nfmt:7, {{s[0-9]+}} idxen offset:42
|
|
41 ; GFX10: tbuffer_store_format_xyzw v[0:3], [[ZEROREG]], s[0:3], format:117, {{s[0-9]+}} idxen offset:42
|
|
42 define amdgpu_ps void @tbuffer_store_scalar_and_imm_offs(<4 x i32> inreg, <4 x float> %vdata, i32 inreg %soffset) {
|
|
43 main_body:
|
|
44 %in1 = bitcast <4 x float> %vdata to <4 x i32>
|
|
45 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 42, i32 %soffset, i32 117, i32 0)
|
|
46 ret void
|
|
47 }
|
|
48
|
|
49 ; GCN-LABEL: {{^}}buffer_store_idx:
|
|
50 ; PREGFX10: tbuffer_store_format_xyzw v[0:3], v4, s[0:3], dfmt:15, nfmt:2, 0 idxen
|
|
51 ; GFX10: tbuffer_store_format_xyzw v[0:3], v4, s[0:3], format:47, 0 idxen
|
|
52 define amdgpu_ps void @buffer_store_idx(<4 x i32> inreg, <4 x float> %vdata, i32 %vindex) {
|
|
53 main_body:
|
|
54 %in1 = bitcast <4 x float> %vdata to <4 x i32>
|
|
55 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 %vindex, i32 0, i32 0, i32 47, i32 0)
|
|
56 ret void
|
|
57 }
|
|
58
|
|
59 ; GCN-LABEL: {{^}}buffer_store_ofs:
|
|
60 ; PREGFX10: tbuffer_store_format_xyzw v[0:3], {{v\[[0-9]+:[0-9]+\]}}, s[0:3], dfmt:3, nfmt:7, 0 idxen offen
|
|
61 ; GFX10: tbuffer_store_format_xyzw v[0:3], {{v\[[0-9]+:[0-9]+\]}}, s[0:3], format:115, 0 idxen offen
|
|
62 define amdgpu_ps void @buffer_store_ofs(<4 x i32> inreg, <4 x float> %vdata, i32 %voffset) {
|
|
63 main_body:
|
|
64 %in1 = bitcast <4 x float> %vdata to <4 x i32>
|
|
65 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 0, i32 %voffset, i32 0, i32 115, i32 0)
|
|
66 ret void
|
|
67 }
|
|
68
|
|
69 ; GCN-LABEL: {{^}}buffer_store_both:
|
|
70 ; PREGFX10: tbuffer_store_format_xyzw v[0:3], v[4:5], s[0:3], dfmt:6, nfmt:4, 0 idxen offen
|
|
71 ; GFX10: tbuffer_store_format_xyzw v[0:3], v[4:5], s[0:3], format:70, 0 idxen offen
|
|
72 define amdgpu_ps void @buffer_store_both(<4 x i32> inreg, <4 x float> %vdata, i32 %vindex, i32 %voffset) {
|
|
73 main_body:
|
|
74 %in1 = bitcast <4 x float> %vdata to <4 x i32>
|
|
75 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 %vindex, i32 %voffset, i32 0, i32 70, i32 0)
|
|
76 ret void
|
|
77 }
|
|
78
|
|
79 ; Ideally, the register allocator would avoid the wait here
|
|
80 ;
|
|
81 ; GCN-LABEL: {{^}}buffer_store_wait:
|
|
82 ; PREGFX10: tbuffer_store_format_xyzw v[0:3], v4, s[0:3], dfmt:15, nfmt:3, 0 idxen
|
|
83 ; GFX10: tbuffer_store_format_xyzw v[0:3], v4, s[0:3], format:63, 0 idxen
|
|
84 ; VERDE: s_waitcnt expcnt(0)
|
|
85 ; GCN: buffer_load_format_xyzw v[0:3], v5, s[0:3], 0 idxen
|
|
86 ; GCN: s_waitcnt vmcnt(0)
|
|
87 ; PREGFX10: tbuffer_store_format_xyzw v[0:3], v6, s[0:3], dfmt:14, nfmt:2, 0 idxen
|
|
88 ; GFX10: tbuffer_store_format_xyzw v[0:3], v6, s[0:3], format:46, 0 idxen
|
|
89 define amdgpu_ps void @buffer_store_wait(<4 x i32> inreg, <4 x float> %vdata, i32 %vindex.1, i32 %vindex.2, i32 %vindex.3) {
|
|
90 main_body:
|
|
91 %in1 = bitcast <4 x float> %vdata to <4 x i32>
|
|
92 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %in1, <4 x i32> %0, i32 %vindex.1, i32 0, i32 0, i32 63, i32 0)
|
|
93 %data = call <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32> %0, i32 %vindex.2, i32 0, i1 0, i1 0)
|
|
94 %data.i = bitcast <4 x float> %data to <4 x i32>
|
|
95 call void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32> %data.i, <4 x i32> %0, i32 %vindex.3, i32 0, i32 0, i32 46, i32 0)
|
|
96 ret void
|
|
97 }
|
|
98
|
|
99 ; GCN-LABEL: {{^}}buffer_store_x1:
|
|
100 ; PREGFX10: tbuffer_store_format_x v0, v1, s[0:3], dfmt:13, nfmt:7, 0 idxen
|
|
101 ; GFX10: tbuffer_store_format_x v0, v1, s[0:3], format:125, 0 idxen
|
|
102 define amdgpu_ps void @buffer_store_x1(<4 x i32> inreg %rsrc, float %data, i32 %vindex) {
|
|
103 main_body:
|
|
104 %data.i = bitcast float %data to i32
|
|
105 call void @llvm.amdgcn.struct.tbuffer.store.i32(i32 %data.i, <4 x i32> %rsrc, i32 %vindex, i32 0, i32 0, i32 125, i32 0)
|
|
106 ret void
|
|
107 }
|
|
108
|
|
109 ; GCN-LABEL: {{^}}buffer_store_x2:
|
|
110 ; PREGFX10: tbuffer_store_format_xy v[0:1], v2, s[0:3], dfmt:1, nfmt:2, 0 idxen
|
|
111 ; GFX10: tbuffer_store_format_xy v[0:1], v2, s[0:3], format:33, 0 idxen
|
|
112 define amdgpu_ps void @buffer_store_x2(<4 x i32> inreg %rsrc, <2 x float> %data, i32 %vindex) {
|
|
113 main_body:
|
|
114 %data.i = bitcast <2 x float> %data to <2 x i32>
|
|
115 call void @llvm.amdgcn.struct.tbuffer.store.v2i32(<2 x i32> %data.i, <4 x i32> %rsrc, i32 %vindex, i32 0, i32 0, i32 33, i32 0)
|
|
116 ret void
|
|
117 }
|
|
118
|
|
119 declare void @llvm.amdgcn.struct.tbuffer.store.i32(i32, <4 x i32>, i32, i32, i32, i32, i32) #0
|
|
120 declare void @llvm.amdgcn.struct.tbuffer.store.v2i32(<2 x i32>, <4 x i32>, i32, i32, i32, i32, i32) #0
|
|
121 declare void @llvm.amdgcn.struct.tbuffer.store.v4i32(<4 x i32>, <4 x i32>, i32, i32, i32, i32, i32) #0
|
|
122 declare void @llvm.amdgcn.struct.tbuffer.store.v4f32(<4 x float>, <4 x i32>, i32, i32, i32, i32, i32) #0
|
|
123 declare <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32>, i32, i32, i1, i1) #1
|
|
124
|
|
125 attributes #0 = { nounwind }
|
|
126 attributes #1 = { nounwind readonly }
|
|
127
|
|
128
|