1; RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs -show-mc-encoding | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=UNPACKED %s
2; RUN: llc < %s -march=amdgcn -mcpu=gfx810 -verify-machineinstrs | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=PACKED %s
3; RUN: llc < %s -march=amdgcn -mcpu=gfx900 -verify-machineinstrs | FileCheck -enable-var-scope -check-prefix=GCN -check-prefix=PACKED %s
4
5; GCN-LABEL: {{^}}tbuffer_load_d16_x:
6; GCN: tbuffer_load_format_d16_x v{{[0-9]+}}, off, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_DATA_FORMAT_10_11_11,BUF_NUM_FORMAT_SNORM]
7define amdgpu_ps half @tbuffer_load_d16_x(<4 x i32> inreg %rsrc) {
8main_body:
9  %data = call half @llvm.amdgcn.tbuffer.load.f16(<4 x i32> %rsrc, i32 0, i32 0, i32 0, i32 0, i32 6, i32 1, i1 0, i1 0)
10  ret half %data
11}
12
13; GCN-LABEL: {{^}}tbuffer_load_d16_xy:
14; UNPACKED: tbuffer_load_format_d16_xy v{{\[}}{{[0-9]+}}:[[HI:[0-9]+]]{{\]}}, off, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_DATA_FORMAT_10_11_11,BUF_NUM_FORMAT_SNORM]
15; UNPACKED: v_mov_b32_e32 v{{[0-9]+}}, v[[HI]]
16
17; PACKED: tbuffer_load_format_d16_xy v[[FULL:[0-9]+]], off, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_DATA_FORMAT_10_11_11,BUF_NUM_FORMAT_SNORM]
18; PACKED: v_lshrrev_b32_e32 v{{[0-9]+}}, 16, v[[FULL]]
19define amdgpu_ps half @tbuffer_load_d16_xy(<4 x i32> inreg %rsrc) {
20main_body:
21  %data = call <2 x half> @llvm.amdgcn.tbuffer.load.v2f16(<4 x i32> %rsrc, i32 0, i32 0, i32 0, i32 0, i32 6, i32 1, i1 0, i1 0)
22  %elt = extractelement <2 x half> %data, i32 1
23  ret half %elt
24}
25
26; GCN-LABEL: {{^}}tbuffer_load_d16_xyz:
27; UNPACKED: tbuffer_load_format_d16_xyz v{{\[}}{{[0-9]+}}:[[HI:[0-9]+]]{{\]}}, off, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_DATA_FORMAT_10_11_11,BUF_NUM_FORMAT_SNORM]
28; UNPACKED: v_mov_b32_e32 v{{[0-9]+}}, v[[HI]]
29
30; PACKED: tbuffer_load_format_d16_xyz v{{\[}}{{[0-9]+}}:[[HI:[0-9]+]]{{\]}}, off, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_DATA_FORMAT_10_11_11,BUF_NUM_FORMAT_SNORM]
31; PACKED: v_mov_b32_e32 v{{[0-9]+}}, v[[HI]]
32define amdgpu_ps half @tbuffer_load_d16_xyz(<4 x i32> inreg %rsrc) {
33main_body:
34  %data = call <3 x half> @llvm.amdgcn.tbuffer.load.v3f16(<4 x i32> %rsrc, i32 0, i32 0, i32 0, i32 0, i32 6, i32 1, i1 0, i1 0)
35  %elt = extractelement <3 x half> %data, i32 2
36  ret half %elt
37}
38
39; GCN-LABEL: {{^}}tbuffer_load_d16_xyzw:
40; UNPACKED: tbuffer_load_format_d16_xyzw v{{\[}}{{[0-9]+}}:[[HI:[0-9]+]]{{\]}}, off, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_DATA_FORMAT_10_11_11,BUF_NUM_FORMAT_SNORM]
41; UNPACKED: v_mov_b32_e32 v{{[0-9]+}}, v[[HI]]
42
43; PACKED: tbuffer_load_format_d16_xyzw v{{\[}}{{[0-9]+}}:[[HI:[0-9]+]]{{\]}}, off, s[{{[0-9]+:[0-9]+}}], 0 format:[BUF_DATA_FORMAT_10_11_11,BUF_NUM_FORMAT_SNORM]
44; PACKED: v_lshrrev_b32_e32 v{{[0-9]+}}, 16, v[[HI]]
45define amdgpu_ps half @tbuffer_load_d16_xyzw(<4 x i32> inreg %rsrc) {
46main_body:
47  %data = call <4 x half> @llvm.amdgcn.tbuffer.load.v4f16(<4 x i32> %rsrc, i32 0, i32 0, i32 0, i32 0, i32 6, i32 1, i1 0, i1 0)
48  %elt = extractelement <4 x half> %data, i32 3
49  ret half %elt
50}
51
52declare half @llvm.amdgcn.tbuffer.load.f16(<4 x i32>, i32, i32, i32, i32, i32, i32, i1, i1)
53declare <2 x half> @llvm.amdgcn.tbuffer.load.v2f16(<4 x i32>, i32, i32, i32, i32, i32, i32, i1, i1)
54declare <3 x half> @llvm.amdgcn.tbuffer.load.v3f16(<4 x i32>, i32, i32, i32, i32, i32, i32, i1, i1)
55declare <4 x half> @llvm.amdgcn.tbuffer.load.v4f16(<4 x i32>, i32, i32, i32, i32, i32, i32, i1, i1)
56