150
|
1 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=tonga -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,PREGFX10,UNPACKED,PREGFX10-UNPACKED %s
|
221
|
2 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx810 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,PREGFX10,PACKED,PREGFX10-PACKED %s
|
|
3 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,PREGFX10,PACKED,PREGFX10-PACKED %s
|
150
|
4 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1010 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX10,PACKED,GFX10-PACKED %s
|
|
5
|
|
6
|
|
7 ; GCN-LABEL: {{^}}tbuffer_store_d16_x:
|
|
8 ; GCN-DAG: s_load_dwordx4
|
|
9 ; GCN-DAG: s_load_dword{{[x0-2]*}} s{{\[}}[[S_LO:[0-9]+]]
|
|
10 ; GCN-DAG: v_mov_b32_e32 v[[V_LO:[0-9]+]], s[[S_LO]]
|
221
|
11 ; PREGFX10: tbuffer_store_format_d16_x v[[V_LO]], v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_NUM_FORMAT_USCALED] idxen
|
|
12 ; GFX10: tbuffer_store_format_d16_x v[[V_LO]], v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_FMT_10_11_11_SSCALED] idxen
|
150
|
13 define amdgpu_kernel void @tbuffer_store_d16_x(<4 x i32> %rsrc, half %data, i32 %vindex) {
|
|
14 main_body:
|
|
15 call void @llvm.amdgcn.struct.tbuffer.store.f16(half %data, <4 x i32> %rsrc, i32 %vindex, i32 0, i32 0, i32 33, i32 0)
|
|
16 ret void
|
|
17 }
|
|
18
|
|
19 ; GCN-LABEL: {{^}}tbuffer_store_d16_xy:
|
|
20 ; GCN: s_load_dword [[S_DATA:s[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0x10
|
|
21 ; UNPACKED-DAG: s_lshr_b32 [[SHR:s[0-9]+]], [[S_DATA]], 16
|
|
22 ; UNPACKED-DAG: s_and_b32 [[MASKED:s[0-9]+]], [[S_DATA]], 0xffff{{$}}
|
|
23 ; UNPACKED-DAG: v_mov_b32_e32 v[[V_LO:[0-9]+]], [[MASKED]]
|
|
24 ; UNPACKED-DAG: v_mov_b32_e32 v[[V_HI:[0-9]+]], [[SHR]]
|
221
|
25 ; PREGFX10-UNPACKED: tbuffer_store_format_d16_xy v{{\[}}[[V_LO]]:[[V_HI]]{{\]}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_NUM_FORMAT_USCALED] idxen
|
150
|
26
|
221
|
27 ; PREGFX10-PACKED: tbuffer_store_format_d16_xy v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_NUM_FORMAT_USCALED] idxen
|
|
28 ; GFX10-PACKED: tbuffer_store_format_d16_xy v{{[0-9]+}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_FMT_10_11_11_SSCALED] idxen
|
150
|
29 define amdgpu_kernel void @tbuffer_store_d16_xy(<4 x i32> %rsrc, <2 x half> %data, i32 %vindex) {
|
|
30 main_body:
|
|
31 call void @llvm.amdgcn.struct.tbuffer.store.v2f16(<2 x half> %data, <4 x i32> %rsrc, i32 %vindex, i32 0, i32 0, i32 33, i32 0)
|
|
32 ret void
|
|
33 }
|
|
34
|
221
|
35 ; GCN-LABEL: {{^}}tbuffer_store_d16_xyz:
|
|
36 ; GCN-DAG: s_load_dwordx2 s{{\[}}[[S_DATA_0:[0-9]+]]:[[S_DATA_1:[0-9]+]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0x10
|
|
37
|
|
38 ; UNPACKED-DAG: s_mov_b32 [[K:s[0-9]+]], 0xffff{{$}}
|
|
39 ; UNPACKED-DAG: s_lshr_b32 [[SHR0:s[0-9]+]], s[[S_DATA_0]], 16
|
|
40 ; UNPACKED-DAG: s_and_b32 [[MASKED0:s[0-9]+]], s[[S_DATA_0]], [[K]]
|
|
41 ; UNPACKED-DAG: s_and_b32 [[MASKED1:s[0-9]+]], s[[S_DATA_1]], [[K]]
|
|
42
|
|
43 ; UNPACKED-DAG: v_mov_b32_e32 v[[LO:[0-9]+]], [[MASKED0]]
|
|
44 ; UNPACKED-DAG: v_mov_b32_e32 v[[HI:[0-9]+]], [[MASKED1]]
|
|
45 ; PREGFX10-UNPACKED: tbuffer_store_format_d16_xyz v{{\[}}[[LO]]:[[HI]]{{\]}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_NUM_FORMAT_USCALED] idxen
|
|
46
|
|
47 ; PACKED-DAG: s_and_b32 [[MASKED0:s[0-9]+]], s[[S_DATA_1]], 0xffff{{$}}
|
|
48 ; PACKED-DAG: v_mov_b32_e32 v[[LO:[0-9]+]], s[[S_DATA_0]]
|
|
49 ; PACKED-DAG: v_mov_b32_e32 v[[HI:[0-9]+]], [[MASKED0]]
|
|
50 ; PREGFX10-PACKED: tbuffer_store_format_d16_xyz v{{\[}}[[LO]]:[[HI]]{{\]}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_NUM_FORMAT_USCALED] idxen
|
|
51 ; GFX10-PACKED: tbuffer_store_format_d16_xyz v{{\[}}[[LO]]:[[HI]]{{\]}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_FMT_10_11_11_SSCALED] idxen
|
|
52 define amdgpu_kernel void @tbuffer_store_d16_xyz(<4 x i32> %rsrc, <4 x half> %data, i32 %vindex) {
|
|
53 main_body:
|
|
54 %data_subvec = shufflevector <4 x half> %data, <4 x half> undef, <3 x i32> <i32 0, i32 1, i32 2>
|
|
55 call void @llvm.amdgcn.struct.tbuffer.store.v3f16(<3 x half> %data_subvec, <4 x i32> %rsrc, i32 %vindex, i32 0, i32 0, i32 33, i32 0)
|
|
56 ret void
|
|
57 }
|
|
58
|
150
|
59 ; GCN-LABEL: {{^}}tbuffer_store_d16_xyzw:
|
|
60 ; GCN-DAG: s_load_dwordx2 s{{\[}}[[S_DATA_0:[0-9]+]]:[[S_DATA_1:[0-9]+]]{{\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0x10
|
|
61
|
|
62 ; UNPACKED-DAG: s_mov_b32 [[K:s[0-9]+]], 0xffff{{$}}
|
|
63 ; UNPACKED-DAG: s_lshr_b32 [[SHR0:s[0-9]+]], s[[S_DATA_0]], 16
|
|
64 ; UNPACKED-DAG: s_and_b32 [[MASKED0:s[0-9]+]], s[[S_DATA_0]], [[K]]
|
|
65 ; UNPACKED-DAG: s_lshr_b32 [[SHR1:s[0-9]+]], s[[S_DATA_1]], 16
|
|
66 ; UNPACKED-DAG: s_and_b32 [[MASKED1:s[0-9]+]], s[[S_DATA_1]], [[K]]
|
|
67
|
|
68 ; UNPACKED-DAG: v_mov_b32_e32 v[[LO:[0-9]+]], [[MASKED0]]
|
|
69 ; UNPACKED-DAG: v_mov_b32_e32 v[[HI:[0-9]+]], [[SHR1]]
|
221
|
70 ; PREGFX10-UNPACKED: tbuffer_store_format_d16_xyzw v{{\[}}[[LO]]:[[HI]]{{\]}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_NUM_FORMAT_USCALED] idxen
|
150
|
71
|
|
72 ; PACKED-DAG: v_mov_b32_e32 v[[LO:[0-9]+]], s[[S_DATA_0]]
|
|
73 ; PACKED-DAG: v_mov_b32_e32 v[[HI:[0-9]+]], s[[S_DATA_1]]
|
221
|
74 ; PREGFX10-PACKED: tbuffer_store_format_d16_xyzw v{{\[}}[[LO]]:[[HI]]{{\]}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_NUM_FORMAT_USCALED] idxen
|
|
75 ; GFX10-PACKED: tbuffer_store_format_d16_xyzw v{{\[}}[[LO]]:[[HI]]{{\]}}, v{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_FMT_10_11_11_SSCALED] idxen
|
150
|
76 define amdgpu_kernel void @tbuffer_store_d16_xyzw(<4 x i32> %rsrc, <4 x half> %data, i32 %vindex) {
|
|
77 main_body:
|
|
78 call void @llvm.amdgcn.struct.tbuffer.store.v4f16(<4 x half> %data, <4 x i32> %rsrc, i32 %vindex, i32 0, i32 0, i32 33, i32 0)
|
|
79 ret void
|
|
80 }
|
|
81
|
|
82 declare void @llvm.amdgcn.struct.tbuffer.store.f16(half, <4 x i32>, i32, i32, i32, i32, i32)
|
|
83 declare void @llvm.amdgcn.struct.tbuffer.store.v2f16(<2 x half>, <4 x i32>, i32, i32, i32, i32, i32)
|
221
|
84 declare void @llvm.amdgcn.struct.tbuffer.store.v3f16(<3 x half>, <4 x i32>, i32, i32, i32, i32, i32)
|
150
|
85 declare void @llvm.amdgcn.struct.tbuffer.store.v4f16(<4 x half>, <4 x i32>, i32, i32, i32, i32, i32)
|