xref: /llvm-project/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.image.a16.encode.ll (revision 3277c7cd28154e33637a168acb26cea7ac1f7fff) 
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc -mtriple=amdgcn -mcpu=gfx900 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX9 %s
3; RUN: llc -mtriple=amdgcn -mcpu=gfx1010 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX10 %s
4; RUN: llc -mtriple=amdgcn -mcpu=gfx1100 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX11 %s
5; RUN: llc -mtriple=amdgcn -mcpu=gfx1200 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX12 %s
6
7define amdgpu_ps <4 x float> @load_1d(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
8; GFX9-LABEL: load_1d:
9; GFX9:       ; %bb.0: ; %main_body
10; GFX9-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x00,0xf0,0x00,0x00,0x00,0x00]
11; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
12; GFX9-NEXT:    ; return to shader part epilog
13;
14; GFX10-LABEL: load_1d:
15; GFX10:       ; %bb.0: ; %main_body
16; GFX10-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
17; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
18; GFX10-NEXT:    ; return to shader part epilog
19;
20; GFX11-LABEL: load_1d:
21; GFX11:       ; %bb.0: ; %main_body
22; GFX11-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x0f,0x01,0xf0,0x00,0x00,0x00,0x00]
23; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
24; GFX11-NEXT:    ; return to shader part epilog
25;
26; GFX12-LABEL: load_1d:
27; GFX12:       ; %bb.0: ; %main_body
28; GFX12-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0x00,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
29; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
30; GFX12-NEXT:    ; return to shader part epilog
31main_body:
32  %s = extractelement <2 x i16> %coords, i32 0
33  %v = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i16(i32 15, i16 %s, <8 x i32> %rsrc, i32 0, i32 0)
34  ret <4 x float> %v
35}
36
37define amdgpu_ps <4 x float> @load_2d(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
38; GFX9-LABEL: load_2d:
39; GFX9:       ; %bb.0: ; %main_body
40; GFX9-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x00,0xf0,0x00,0x00,0x00,0x00]
41; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
42; GFX9-NEXT:    ; return to shader part epilog
43;
44; GFX10-LABEL: load_2d:
45; GFX10:       ; %bb.0: ; %main_body
46; GFX10-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x08,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
47; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
48; GFX10-NEXT:    ; return to shader part epilog
49;
50; GFX11-LABEL: load_2d:
51; GFX11:       ; %bb.0: ; %main_body
52; GFX11-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x84,0x0f,0x01,0xf0,0x00,0x00,0x00,0x00]
53; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
54; GFX11-NEXT:    ; return to shader part epilog
55;
56; GFX12-LABEL: load_2d:
57; GFX12:       ; %bb.0: ; %main_body
58; GFX12-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x41,0x00,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
59; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
60; GFX12-NEXT:    ; return to shader part epilog
61main_body:
62  %s = extractelement <2 x i16> %coords, i32 0
63  %t = extractelement <2 x i16> %coords, i32 1
64  %v = call <4 x float> @llvm.amdgcn.image.load.2d.v4f32.i16(i32 15, i16 %s, i16 %t, <8 x i32> %rsrc, i32 0, i32 0)
65  ret <4 x float> %v
66}
67
68define amdgpu_ps <4 x float> @load_3d(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
69; GFX9-LABEL: load_3d:
70; GFX9:       ; %bb.0: ; %main_body
71; GFX9-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x00,0xf0,0x00,0x00,0x00,0x00]
72; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
73; GFX9-NEXT:    ; return to shader part epilog
74;
75; GFX10-LABEL: load_3d:
76; GFX10:       ; %bb.0: ; %main_body
77; GFX10-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x10,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
78; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
79; GFX10-NEXT:    ; return to shader part epilog
80;
81; GFX11-LABEL: load_3d:
82; GFX11:       ; %bb.0: ; %main_body
83; GFX11-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x88,0x0f,0x01,0xf0,0x00,0x00,0x00,0x00]
84; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
85; GFX11-NEXT:    ; return to shader part epilog
86;
87; GFX12-LABEL: load_3d:
88; GFX12:       ; %bb.0: ; %main_body
89; GFX12-NEXT:    image_load v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D a16 ; encoding: [0x42,0x00,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
90; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
91; GFX12-NEXT:    ; return to shader part epilog
92main_body:
93  %s = extractelement <2 x i16> %coords_lo, i32 0
94  %t = extractelement <2 x i16> %coords_lo, i32 1
95  %r = extractelement <2 x i16> %coords_hi, i32 0
96  %v = call <4 x float> @llvm.amdgcn.image.load.3d.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %r, <8 x i32> %rsrc, i32 0, i32 0)
97  ret <4 x float> %v
98}
99
100define amdgpu_ps <4 x float> @load_cube(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
101; GFX9-LABEL: load_cube:
102; GFX9:       ; %bb.0: ; %main_body
103; GFX9-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x00,0xf0,0x00,0x00,0x00,0x00]
104; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
105; GFX9-NEXT:    ; return to shader part epilog
106;
107; GFX10-LABEL: load_cube:
108; GFX10:       ; %bb.0: ; %main_body
109; GFX10-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x18,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
110; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
111; GFX10-NEXT:    ; return to shader part epilog
112;
113; GFX11-LABEL: load_cube:
114; GFX11:       ; %bb.0: ; %main_body
115; GFX11-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x8c,0x0f,0x01,0xf0,0x00,0x00,0x00,0x00]
116; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
117; GFX11-NEXT:    ; return to shader part epilog
118;
119; GFX12-LABEL: load_cube:
120; GFX12:       ; %bb.0: ; %main_body
121; GFX12-NEXT:    image_load v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE a16 ; encoding: [0x43,0x00,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
122; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
123; GFX12-NEXT:    ; return to shader part epilog
124main_body:
125  %s = extractelement <2 x i16> %coords_lo, i32 0
126  %t = extractelement <2 x i16> %coords_lo, i32 1
127  %slice = extractelement <2 x i16> %coords_hi, i32 0
128  %v = call <4 x float> @llvm.amdgcn.image.load.cube.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %slice, <8 x i32> %rsrc, i32 0, i32 0)
129  ret <4 x float> %v
130}
131
132define amdgpu_ps <4 x float> @load_1darray(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
133; GFX9-LABEL: load_1darray:
134; GFX9:       ; %bb.0: ; %main_body
135; GFX9-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x00,0xf0,0x00,0x00,0x00,0x00]
136; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
137; GFX9-NEXT:    ; return to shader part epilog
138;
139; GFX10-LABEL: load_1darray:
140; GFX10:       ; %bb.0: ; %main_body
141; GFX10-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x20,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
142; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
143; GFX10-NEXT:    ; return to shader part epilog
144;
145; GFX11-LABEL: load_1darray:
146; GFX11:       ; %bb.0: ; %main_body
147; GFX11-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x90,0x0f,0x01,0xf0,0x00,0x00,0x00,0x00]
148; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
149; GFX11-NEXT:    ; return to shader part epilog
150;
151; GFX12-LABEL: load_1darray:
152; GFX12:       ; %bb.0: ; %main_body
153; GFX12-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY a16 ; encoding: [0x44,0x00,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
154; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
155; GFX12-NEXT:    ; return to shader part epilog
156main_body:
157  %s = extractelement <2 x i16> %coords, i32 0
158  %slice = extractelement <2 x i16> %coords, i32 1
159  %v = call <4 x float> @llvm.amdgcn.image.load.1darray.v4f32.i16(i32 15, i16 %s, i16 %slice, <8 x i32> %rsrc, i32 0, i32 0)
160  ret <4 x float> %v
161}
162
163define amdgpu_ps <4 x float> @load_2darray(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
164; GFX9-LABEL: load_2darray:
165; GFX9:       ; %bb.0: ; %main_body
166; GFX9-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x00,0xf0,0x00,0x00,0x00,0x00]
167; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
168; GFX9-NEXT:    ; return to shader part epilog
169;
170; GFX10-LABEL: load_2darray:
171; GFX10:       ; %bb.0: ; %main_body
172; GFX10-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x28,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
173; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
174; GFX10-NEXT:    ; return to shader part epilog
175;
176; GFX11-LABEL: load_2darray:
177; GFX11:       ; %bb.0: ; %main_body
178; GFX11-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x94,0x0f,0x01,0xf0,0x00,0x00,0x00,0x00]
179; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
180; GFX11-NEXT:    ; return to shader part epilog
181;
182; GFX12-LABEL: load_2darray:
183; GFX12:       ; %bb.0: ; %main_body
184; GFX12-NEXT:    image_load v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x45,0x00,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
185; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
186; GFX12-NEXT:    ; return to shader part epilog
187main_body:
188  %s = extractelement <2 x i16> %coords_lo, i32 0
189  %t = extractelement <2 x i16> %coords_lo, i32 1
190  %slice = extractelement <2 x i16> %coords_hi, i32 0
191  %v = call <4 x float> @llvm.amdgcn.image.load.2darray.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %slice, <8 x i32> %rsrc, i32 0, i32 0)
192  ret <4 x float> %v
193}
194
195define amdgpu_ps <4 x float> @load_2dmsaa(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
196; GFX9-LABEL: load_2dmsaa:
197; GFX9:       ; %bb.0: ; %main_body
198; GFX9-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x00,0xf0,0x00,0x00,0x00,0x00]
199; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
200; GFX9-NEXT:    ; return to shader part epilog
201;
202; GFX10-LABEL: load_2dmsaa:
203; GFX10:       ; %bb.0: ; %main_body
204; GFX10-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm a16 ; encoding: [0x30,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
205; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
206; GFX10-NEXT:    ; return to shader part epilog
207;
208; GFX11-LABEL: load_2dmsaa:
209; GFX11:       ; %bb.0: ; %main_body
210; GFX11-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm a16 ; encoding: [0x98,0x0f,0x01,0xf0,0x00,0x00,0x00,0x00]
211; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
212; GFX11-NEXT:    ; return to shader part epilog
213;
214; GFX12-LABEL: load_2dmsaa:
215; GFX12:       ; %bb.0: ; %main_body
216; GFX12-NEXT:    image_load v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA a16 ; encoding: [0x46,0x00,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
217; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
218; GFX12-NEXT:    ; return to shader part epilog
219main_body:
220  %s = extractelement <2 x i16> %coords_lo, i32 0
221  %t = extractelement <2 x i16> %coords_lo, i32 1
222  %fragid = extractelement <2 x i16> %coords_hi, i32 0
223  %v = call <4 x float> @llvm.amdgcn.image.load.2dmsaa.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
224  ret <4 x float> %v
225}
226
227define amdgpu_ps <4 x float> @load_2darraymsaa(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
228; GFX9-LABEL: load_2darraymsaa:
229; GFX9:       ; %bb.0: ; %main_body
230; GFX9-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x00,0xf0,0x00,0x00,0x00,0x00]
231; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
232; GFX9-NEXT:    ; return to shader part epilog
233;
234; GFX10-LABEL: load_2darraymsaa:
235; GFX10:       ; %bb.0: ; %main_body
236; GFX10-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm a16 ; encoding: [0x38,0x1f,0x00,0xf0,0x00,0x00,0x00,0x40]
237; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
238; GFX10-NEXT:    ; return to shader part epilog
239;
240; GFX11-LABEL: load_2darraymsaa:
241; GFX11:       ; %bb.0: ; %main_body
242; GFX11-NEXT:    image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm a16 ; encoding: [0x9c,0x0f,0x01,0xf0,0x00,0x00,0x00,0x00]
243; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
244; GFX11-NEXT:    ; return to shader part epilog
245;
246; GFX12-LABEL: load_2darraymsaa:
247; GFX12:       ; %bb.0: ; %main_body
248; GFX12-NEXT:    image_load v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY a16 ; encoding: [0x47,0x00,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
249; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
250; GFX12-NEXT:    ; return to shader part epilog
251main_body:
252  %s = extractelement <2 x i16> %coords_lo, i32 0
253  %t = extractelement <2 x i16> %coords_lo, i32 1
254  %slice = extractelement <2 x i16> %coords_hi, i32 0
255  %fragid = extractelement <2 x i16> %coords_hi, i32 1
256  %v = call <4 x float> @llvm.amdgcn.image.load.2darraymsaa.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %slice, i16 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
257  ret <4 x float> %v
258}
259
260define amdgpu_ps <4 x float> @load_mip_1d(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
261; GFX9-LABEL: load_mip_1d:
262; GFX9:       ; %bb.0: ; %main_body
263; GFX9-NEXT:    image_load_mip v[0:3], v0, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x04,0xf0,0x00,0x00,0x00,0x00]
264; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
265; GFX9-NEXT:    ; return to shader part epilog
266;
267; GFX10-LABEL: load_mip_1d:
268; GFX10:       ; %bb.0: ; %main_body
269; GFX10-NEXT:    image_load_mip v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x1f,0x04,0xf0,0x00,0x00,0x00,0x40]
270; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
271; GFX10-NEXT:    ; return to shader part epilog
272;
273; GFX11-LABEL: load_mip_1d:
274; GFX11:       ; %bb.0: ; %main_body
275; GFX11-NEXT:    image_load_mip v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x0f,0x05,0xf0,0x00,0x00,0x00,0x00]
276; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
277; GFX11-NEXT:    ; return to shader part epilog
278;
279; GFX12-LABEL: load_mip_1d:
280; GFX12:       ; %bb.0: ; %main_body
281; GFX12-NEXT:    image_load_mip v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0x40,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
282; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
283; GFX12-NEXT:    ; return to shader part epilog
284main_body:
285  %s = extractelement <2 x i16> %coords, i32 0
286  %mip = extractelement <2 x i16> %coords, i32 1
287  %v = call <4 x float> @llvm.amdgcn.image.load.mip.1d.v4f32.i16(i32 15, i16 %s, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
288  ret <4 x float> %v
289}
290
291define amdgpu_ps <4 x float> @load_mip_2d(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
292; GFX9-LABEL: load_mip_2d:
293; GFX9:       ; %bb.0: ; %main_body
294; GFX9-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x04,0xf0,0x00,0x00,0x00,0x00]
295; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
296; GFX9-NEXT:    ; return to shader part epilog
297;
298; GFX10-LABEL: load_mip_2d:
299; GFX10:       ; %bb.0: ; %main_body
300; GFX10-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x08,0x1f,0x04,0xf0,0x00,0x00,0x00,0x40]
301; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
302; GFX10-NEXT:    ; return to shader part epilog
303;
304; GFX11-LABEL: load_mip_2d:
305; GFX11:       ; %bb.0: ; %main_body
306; GFX11-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x84,0x0f,0x05,0xf0,0x00,0x00,0x00,0x00]
307; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
308; GFX11-NEXT:    ; return to shader part epilog
309;
310; GFX12-LABEL: load_mip_2d:
311; GFX12:       ; %bb.0: ; %main_body
312; GFX12-NEXT:    image_load_mip v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x41,0x40,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
313; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
314; GFX12-NEXT:    ; return to shader part epilog
315main_body:
316  %s = extractelement <2 x i16> %coords_lo, i32 0
317  %t = extractelement <2 x i16> %coords_lo, i32 1
318  %mip = extractelement <2 x i16> %coords_hi, i32 0
319  %v = call <4 x float> @llvm.amdgcn.image.load.mip.2d.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
320  ret <4 x float> %v
321}
322
323define amdgpu_ps <4 x float> @load_mip_3d(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
324; GFX9-LABEL: load_mip_3d:
325; GFX9:       ; %bb.0: ; %main_body
326; GFX9-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x04,0xf0,0x00,0x00,0x00,0x00]
327; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
328; GFX9-NEXT:    ; return to shader part epilog
329;
330; GFX10-LABEL: load_mip_3d:
331; GFX10:       ; %bb.0: ; %main_body
332; GFX10-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x10,0x1f,0x04,0xf0,0x00,0x00,0x00,0x40]
333; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
334; GFX10-NEXT:    ; return to shader part epilog
335;
336; GFX11-LABEL: load_mip_3d:
337; GFX11:       ; %bb.0: ; %main_body
338; GFX11-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x88,0x0f,0x05,0xf0,0x00,0x00,0x00,0x00]
339; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
340; GFX11-NEXT:    ; return to shader part epilog
341;
342; GFX12-LABEL: load_mip_3d:
343; GFX12:       ; %bb.0: ; %main_body
344; GFX12-NEXT:    image_load_mip v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D a16 ; encoding: [0x42,0x40,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
345; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
346; GFX12-NEXT:    ; return to shader part epilog
347main_body:
348  %s = extractelement <2 x i16> %coords_lo, i32 0
349  %t = extractelement <2 x i16> %coords_lo, i32 1
350  %r = extractelement <2 x i16> %coords_hi, i32 0
351  %mip = extractelement <2 x i16> %coords_hi, i32 1
352  %v = call <4 x float> @llvm.amdgcn.image.load.mip.3d.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %r, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
353  ret <4 x float> %v
354}
355
356define amdgpu_ps <4 x float> @load_mip_cube(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
357; GFX9-LABEL: load_mip_cube:
358; GFX9:       ; %bb.0: ; %main_body
359; GFX9-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x04,0xf0,0x00,0x00,0x00,0x00]
360; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
361; GFX9-NEXT:    ; return to shader part epilog
362;
363; GFX10-LABEL: load_mip_cube:
364; GFX10:       ; %bb.0: ; %main_body
365; GFX10-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x18,0x1f,0x04,0xf0,0x00,0x00,0x00,0x40]
366; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
367; GFX10-NEXT:    ; return to shader part epilog
368;
369; GFX11-LABEL: load_mip_cube:
370; GFX11:       ; %bb.0: ; %main_body
371; GFX11-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x8c,0x0f,0x05,0xf0,0x00,0x00,0x00,0x00]
372; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
373; GFX11-NEXT:    ; return to shader part epilog
374;
375; GFX12-LABEL: load_mip_cube:
376; GFX12:       ; %bb.0: ; %main_body
377; GFX12-NEXT:    image_load_mip v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE a16 ; encoding: [0x43,0x40,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
378; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
379; GFX12-NEXT:    ; return to shader part epilog
380main_body:
381  %s = extractelement <2 x i16> %coords_lo, i32 0
382  %t = extractelement <2 x i16> %coords_lo, i32 1
383  %slice = extractelement <2 x i16> %coords_hi, i32 0
384  %mip = extractelement <2 x i16> %coords_hi, i32 1
385  %v = call <4 x float> @llvm.amdgcn.image.load.mip.cube.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %slice, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
386  ret <4 x float> %v
387}
388
389define amdgpu_ps <4 x float> @load_mip_1darray(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
390; GFX9-LABEL: load_mip_1darray:
391; GFX9:       ; %bb.0: ; %main_body
392; GFX9-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x04,0xf0,0x00,0x00,0x00,0x00]
393; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
394; GFX9-NEXT:    ; return to shader part epilog
395;
396; GFX10-LABEL: load_mip_1darray:
397; GFX10:       ; %bb.0: ; %main_body
398; GFX10-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x20,0x1f,0x04,0xf0,0x00,0x00,0x00,0x40]
399; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
400; GFX10-NEXT:    ; return to shader part epilog
401;
402; GFX11-LABEL: load_mip_1darray:
403; GFX11:       ; %bb.0: ; %main_body
404; GFX11-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x90,0x0f,0x05,0xf0,0x00,0x00,0x00,0x00]
405; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
406; GFX11-NEXT:    ; return to shader part epilog
407;
408; GFX12-LABEL: load_mip_1darray:
409; GFX12:       ; %bb.0: ; %main_body
410; GFX12-NEXT:    image_load_mip v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY a16 ; encoding: [0x44,0x40,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
411; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
412; GFX12-NEXT:    ; return to shader part epilog
413main_body:
414  %s = extractelement <2 x i16> %coords_lo, i32 0
415  %slice = extractelement <2 x i16> %coords_lo, i32 1
416  %mip = extractelement <2 x i16> %coords_hi, i32 0
417  %v = call <4 x float> @llvm.amdgcn.image.load.mip.1darray.v4f32.i16(i32 15, i16 %s, i16 %slice, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
418  ret <4 x float> %v
419}
420
421define amdgpu_ps <4 x float> @load_mip_2darray(<8 x i32> inreg %rsrc, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
422; GFX9-LABEL: load_mip_2darray:
423; GFX9:       ; %bb.0: ; %main_body
424; GFX9-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x04,0xf0,0x00,0x00,0x00,0x00]
425; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
426; GFX9-NEXT:    ; return to shader part epilog
427;
428; GFX10-LABEL: load_mip_2darray:
429; GFX10:       ; %bb.0: ; %main_body
430; GFX10-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x28,0x1f,0x04,0xf0,0x00,0x00,0x00,0x40]
431; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
432; GFX10-NEXT:    ; return to shader part epilog
433;
434; GFX11-LABEL: load_mip_2darray:
435; GFX11:       ; %bb.0: ; %main_body
436; GFX11-NEXT:    image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x94,0x0f,0x05,0xf0,0x00,0x00,0x00,0x00]
437; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
438; GFX11-NEXT:    ; return to shader part epilog
439;
440; GFX12-LABEL: load_mip_2darray:
441; GFX12:       ; %bb.0: ; %main_body
442; GFX12-NEXT:    image_load_mip v[0:3], [v0, v1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x45,0x40,0xc0,0xd3,0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00]
443; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
444; GFX12-NEXT:    ; return to shader part epilog
445main_body:
446  %s = extractelement <2 x i16> %coords_lo, i32 0
447  %t = extractelement <2 x i16> %coords_lo, i32 1
448  %slice = extractelement <2 x i16> %coords_hi, i32 0
449  %mip = extractelement <2 x i16> %coords_hi, i32 1
450  %v = call <4 x float> @llvm.amdgcn.image.load.mip.2darray.v4f32.i16(i32 15, i16 %s, i16 %t, i16 %slice, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
451  ret <4 x float> %v
452}
453
454define amdgpu_ps void @store_1d(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords) {
455; GFX9-LABEL: store_1d:
456; GFX9:       ; %bb.0: ; %main_body
457; GFX9-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x20,0xf0,0x04,0x00,0x00,0x00]
458; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
459;
460; GFX10-LABEL: store_1d:
461; GFX10:       ; %bb.0: ; %main_body
462; GFX10-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x1f,0x20,0xf0,0x04,0x00,0x00,0x40]
463; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
464;
465; GFX11-LABEL: store_1d:
466; GFX11:       ; %bb.0: ; %main_body
467; GFX11-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x0f,0x19,0xf0,0x04,0x00,0x00,0x00]
468; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
469;
470; GFX12-LABEL: store_1d:
471; GFX12:       ; %bb.0: ; %main_body
472; GFX12-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0x80,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x00,0x00,0x00]
473; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
474main_body:
475  %s = extractelement <2 x i16> %coords, i32 0
476  call void @llvm.amdgcn.image.store.1d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, <8 x i32> %rsrc, i32 0, i32 0)
477  ret void
478}
479
480define amdgpu_ps void @store_2d(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords) {
481; GFX9-LABEL: store_2d:
482; GFX9:       ; %bb.0: ; %main_body
483; GFX9-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x20,0xf0,0x04,0x00,0x00,0x00]
484; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
485;
486; GFX10-LABEL: store_2d:
487; GFX10:       ; %bb.0: ; %main_body
488; GFX10-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x08,0x1f,0x20,0xf0,0x04,0x00,0x00,0x40]
489; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
490;
491; GFX11-LABEL: store_2d:
492; GFX11:       ; %bb.0: ; %main_body
493; GFX11-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x84,0x0f,0x19,0xf0,0x04,0x00,0x00,0x00]
494; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
495;
496; GFX12-LABEL: store_2d:
497; GFX12:       ; %bb.0: ; %main_body
498; GFX12-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x41,0x80,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x00,0x00,0x00]
499; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
500main_body:
501  %s = extractelement <2 x i16> %coords, i32 0
502  %t = extractelement <2 x i16> %coords, i32 1
503  call void @llvm.amdgcn.image.store.2d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, <8 x i32> %rsrc, i32 0, i32 0)
504  ret void
505}
506
507define amdgpu_ps void @store_3d(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
508; GFX9-LABEL: store_3d:
509; GFX9:       ; %bb.0: ; %main_body
510; GFX9-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x20,0xf0,0x04,0x00,0x00,0x00]
511; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
512;
513; GFX10-LABEL: store_3d:
514; GFX10:       ; %bb.0: ; %main_body
515; GFX10-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x10,0x1f,0x20,0xf0,0x04,0x00,0x00,0x40]
516; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
517;
518; GFX11-LABEL: store_3d:
519; GFX11:       ; %bb.0: ; %main_body
520; GFX11-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x88,0x0f,0x19,0xf0,0x04,0x00,0x00,0x00]
521; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
522;
523; GFX12-LABEL: store_3d:
524; GFX12:       ; %bb.0: ; %main_body
525; GFX12-NEXT:    image_store v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D a16 ; encoding: [0x42,0x80,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
526; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
527main_body:
528  %s = extractelement <2 x i16> %coords_lo, i32 0
529  %t = extractelement <2 x i16> %coords_lo, i32 1
530  %r = extractelement <2 x i16> %coords_hi, i32 0
531  call void @llvm.amdgcn.image.store.3d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %r, <8 x i32> %rsrc, i32 0, i32 0)
532  ret void
533}
534
535define amdgpu_ps void @store_cube(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
536; GFX9-LABEL: store_cube:
537; GFX9:       ; %bb.0: ; %main_body
538; GFX9-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x20,0xf0,0x04,0x00,0x00,0x00]
539; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
540;
541; GFX10-LABEL: store_cube:
542; GFX10:       ; %bb.0: ; %main_body
543; GFX10-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x18,0x1f,0x20,0xf0,0x04,0x00,0x00,0x40]
544; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
545;
546; GFX11-LABEL: store_cube:
547; GFX11:       ; %bb.0: ; %main_body
548; GFX11-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x8c,0x0f,0x19,0xf0,0x04,0x00,0x00,0x00]
549; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
550;
551; GFX12-LABEL: store_cube:
552; GFX12:       ; %bb.0: ; %main_body
553; GFX12-NEXT:    image_store v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE a16 ; encoding: [0x43,0x80,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
554; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
555main_body:
556  %s = extractelement <2 x i16> %coords_lo, i32 0
557  %t = extractelement <2 x i16> %coords_lo, i32 1
558  %slice = extractelement <2 x i16> %coords_hi, i32 0
559  call void @llvm.amdgcn.image.store.cube.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %slice, <8 x i32> %rsrc, i32 0, i32 0)
560  ret void
561}
562
563define amdgpu_ps void @store_1darray(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords) {
564; GFX9-LABEL: store_1darray:
565; GFX9:       ; %bb.0: ; %main_body
566; GFX9-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x20,0xf0,0x04,0x00,0x00,0x00]
567; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
568;
569; GFX10-LABEL: store_1darray:
570; GFX10:       ; %bb.0: ; %main_body
571; GFX10-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x20,0x1f,0x20,0xf0,0x04,0x00,0x00,0x40]
572; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
573;
574; GFX11-LABEL: store_1darray:
575; GFX11:       ; %bb.0: ; %main_body
576; GFX11-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x90,0x0f,0x19,0xf0,0x04,0x00,0x00,0x00]
577; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
578;
579; GFX12-LABEL: store_1darray:
580; GFX12:       ; %bb.0: ; %main_body
581; GFX12-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY a16 ; encoding: [0x44,0x80,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x00,0x00,0x00]
582; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
583main_body:
584  %s = extractelement <2 x i16> %coords, i32 0
585  %slice = extractelement <2 x i16> %coords, i32 1
586  call void @llvm.amdgcn.image.store.1darray.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %slice, <8 x i32> %rsrc, i32 0, i32 0)
587  ret void
588}
589
590define amdgpu_ps void @store_2darray(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
591; GFX9-LABEL: store_2darray:
592; GFX9:       ; %bb.0: ; %main_body
593; GFX9-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x20,0xf0,0x04,0x00,0x00,0x00]
594; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
595;
596; GFX10-LABEL: store_2darray:
597; GFX10:       ; %bb.0: ; %main_body
598; GFX10-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x28,0x1f,0x20,0xf0,0x04,0x00,0x00,0x40]
599; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
600;
601; GFX11-LABEL: store_2darray:
602; GFX11:       ; %bb.0: ; %main_body
603; GFX11-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x94,0x0f,0x19,0xf0,0x04,0x00,0x00,0x00]
604; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
605;
606; GFX12-LABEL: store_2darray:
607; GFX12:       ; %bb.0: ; %main_body
608; GFX12-NEXT:    image_store v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x45,0x80,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
609; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
610main_body:
611  %s = extractelement <2 x i16> %coords_lo, i32 0
612  %t = extractelement <2 x i16> %coords_lo, i32 1
613  %slice = extractelement <2 x i16> %coords_hi, i32 0
614  call void @llvm.amdgcn.image.store.2darray.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %slice, <8 x i32> %rsrc, i32 0, i32 0)
615  ret void
616}
617
618define amdgpu_ps void @store_2dmsaa(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
619; GFX9-LABEL: store_2dmsaa:
620; GFX9:       ; %bb.0: ; %main_body
621; GFX9-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x20,0xf0,0x04,0x00,0x00,0x00]
622; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
623;
624; GFX10-LABEL: store_2dmsaa:
625; GFX10:       ; %bb.0: ; %main_body
626; GFX10-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm a16 ; encoding: [0x30,0x1f,0x20,0xf0,0x04,0x00,0x00,0x40]
627; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
628;
629; GFX11-LABEL: store_2dmsaa:
630; GFX11:       ; %bb.0: ; %main_body
631; GFX11-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm a16 ; encoding: [0x98,0x0f,0x19,0xf0,0x04,0x00,0x00,0x00]
632; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
633;
634; GFX12-LABEL: store_2dmsaa:
635; GFX12:       ; %bb.0: ; %main_body
636; GFX12-NEXT:    image_store v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA a16 ; encoding: [0x46,0x80,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
637; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
638main_body:
639  %s = extractelement <2 x i16> %coords_lo, i32 0
640  %t = extractelement <2 x i16> %coords_lo, i32 1
641  %fragid = extractelement <2 x i16> %coords_hi, i32 0
642  call void @llvm.amdgcn.image.store.2dmsaa.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
643  ret void
644}
645
646define amdgpu_ps void @store_2darraymsaa(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
647; GFX9-LABEL: store_2darraymsaa:
648; GFX9:       ; %bb.0: ; %main_body
649; GFX9-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x20,0xf0,0x04,0x00,0x00,0x00]
650; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
651;
652; GFX10-LABEL: store_2darraymsaa:
653; GFX10:       ; %bb.0: ; %main_body
654; GFX10-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm a16 ; encoding: [0x38,0x1f,0x20,0xf0,0x04,0x00,0x00,0x40]
655; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
656;
657; GFX11-LABEL: store_2darraymsaa:
658; GFX11:       ; %bb.0: ; %main_body
659; GFX11-NEXT:    image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm a16 ; encoding: [0x9c,0x0f,0x19,0xf0,0x04,0x00,0x00,0x00]
660; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
661;
662; GFX12-LABEL: store_2darraymsaa:
663; GFX12:       ; %bb.0: ; %main_body
664; GFX12-NEXT:    image_store v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY a16 ; encoding: [0x47,0x80,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
665; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
666main_body:
667  %s = extractelement <2 x i16> %coords_lo, i32 0
668  %t = extractelement <2 x i16> %coords_lo, i32 1
669  %slice = extractelement <2 x i16> %coords_hi, i32 0
670  %fragid = extractelement <2 x i16> %coords_hi, i32 1
671  call void @llvm.amdgcn.image.store.2darraymsaa.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %slice, i16 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
672  ret void
673}
674
675define amdgpu_ps void @store_mip_1d(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords) {
676; GFX9-LABEL: store_mip_1d:
677; GFX9:       ; %bb.0: ; %main_body
678; GFX9-NEXT:    image_store_mip v[0:3], v4, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x24,0xf0,0x04,0x00,0x00,0x00]
679; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
680;
681; GFX10-LABEL: store_mip_1d:
682; GFX10:       ; %bb.0: ; %main_body
683; GFX10-NEXT:    image_store_mip v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x1f,0x24,0xf0,0x04,0x00,0x00,0x40]
684; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
685;
686; GFX11-LABEL: store_mip_1d:
687; GFX11:       ; %bb.0: ; %main_body
688; GFX11-NEXT:    image_store_mip v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x0f,0x1d,0xf0,0x04,0x00,0x00,0x00]
689; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
690;
691; GFX12-LABEL: store_mip_1d:
692; GFX12:       ; %bb.0: ; %main_body
693; GFX12-NEXT:    image_store_mip v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0xc0,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x00,0x00,0x00]
694; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
695main_body:
696  %s = extractelement <2 x i16> %coords, i32 0
697  %mip = extractelement <2 x i16> %coords, i32 1
698  call void @llvm.amdgcn.image.store.mip.1d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
699  ret void
700}
701
702define amdgpu_ps void @store_mip_2d(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
703; GFX9-LABEL: store_mip_2d:
704; GFX9:       ; %bb.0: ; %main_body
705; GFX9-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x24,0xf0,0x04,0x00,0x00,0x00]
706; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
707;
708; GFX10-LABEL: store_mip_2d:
709; GFX10:       ; %bb.0: ; %main_body
710; GFX10-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x08,0x1f,0x24,0xf0,0x04,0x00,0x00,0x40]
711; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
712;
713; GFX11-LABEL: store_mip_2d:
714; GFX11:       ; %bb.0: ; %main_body
715; GFX11-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x84,0x0f,0x1d,0xf0,0x04,0x00,0x00,0x00]
716; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
717;
718; GFX12-LABEL: store_mip_2d:
719; GFX12:       ; %bb.0: ; %main_body
720; GFX12-NEXT:    image_store_mip v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x41,0xc0,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
721; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
722main_body:
723  %s = extractelement <2 x i16> %coords_lo, i32 0
724  %t = extractelement <2 x i16> %coords_lo, i32 1
725  %mip = extractelement <2 x i16> %coords_hi, i32 0
726  call void @llvm.amdgcn.image.store.mip.2d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
727  ret void
728}
729
730define amdgpu_ps void @store_mip_3d(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
731; GFX9-LABEL: store_mip_3d:
732; GFX9:       ; %bb.0: ; %main_body
733; GFX9-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x24,0xf0,0x04,0x00,0x00,0x00]
734; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
735;
736; GFX10-LABEL: store_mip_3d:
737; GFX10:       ; %bb.0: ; %main_body
738; GFX10-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x10,0x1f,0x24,0xf0,0x04,0x00,0x00,0x40]
739; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
740;
741; GFX11-LABEL: store_mip_3d:
742; GFX11:       ; %bb.0: ; %main_body
743; GFX11-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x88,0x0f,0x1d,0xf0,0x04,0x00,0x00,0x00]
744; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
745;
746; GFX12-LABEL: store_mip_3d:
747; GFX12:       ; %bb.0: ; %main_body
748; GFX12-NEXT:    image_store_mip v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D a16 ; encoding: [0x42,0xc0,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
749; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
750main_body:
751  %s = extractelement <2 x i16> %coords_lo, i32 0
752  %t = extractelement <2 x i16> %coords_lo, i32 1
753  %r = extractelement <2 x i16> %coords_hi, i32 0
754  %mip = extractelement <2 x i16> %coords_hi, i32 1
755  call void @llvm.amdgcn.image.store.mip.3d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %r, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
756  ret void
757}
758
759define amdgpu_ps void @store_mip_cube(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
760; GFX9-LABEL: store_mip_cube:
761; GFX9:       ; %bb.0: ; %main_body
762; GFX9-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x24,0xf0,0x04,0x00,0x00,0x00]
763; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
764;
765; GFX10-LABEL: store_mip_cube:
766; GFX10:       ; %bb.0: ; %main_body
767; GFX10-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x18,0x1f,0x24,0xf0,0x04,0x00,0x00,0x40]
768; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
769;
770; GFX11-LABEL: store_mip_cube:
771; GFX11:       ; %bb.0: ; %main_body
772; GFX11-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x8c,0x0f,0x1d,0xf0,0x04,0x00,0x00,0x00]
773; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
774;
775; GFX12-LABEL: store_mip_cube:
776; GFX12:       ; %bb.0: ; %main_body
777; GFX12-NEXT:    image_store_mip v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE a16 ; encoding: [0x43,0xc0,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
778; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
779main_body:
780  %s = extractelement <2 x i16> %coords_lo, i32 0
781  %t = extractelement <2 x i16> %coords_lo, i32 1
782  %slice = extractelement <2 x i16> %coords_hi, i32 0
783  %mip = extractelement <2 x i16> %coords_hi, i32 1
784  call void @llvm.amdgcn.image.store.mip.cube.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %slice, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
785  ret void
786}
787
788define amdgpu_ps void @store_mip_1darray(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
789; GFX9-LABEL: store_mip_1darray:
790; GFX9:       ; %bb.0: ; %main_body
791; GFX9-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x24,0xf0,0x04,0x00,0x00,0x00]
792; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
793;
794; GFX10-LABEL: store_mip_1darray:
795; GFX10:       ; %bb.0: ; %main_body
796; GFX10-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x20,0x1f,0x24,0xf0,0x04,0x00,0x00,0x40]
797; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
798;
799; GFX11-LABEL: store_mip_1darray:
800; GFX11:       ; %bb.0: ; %main_body
801; GFX11-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x90,0x0f,0x1d,0xf0,0x04,0x00,0x00,0x00]
802; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
803;
804; GFX12-LABEL: store_mip_1darray:
805; GFX12:       ; %bb.0: ; %main_body
806; GFX12-NEXT:    image_store_mip v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY a16 ; encoding: [0x44,0xc0,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
807; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
808main_body:
809  %s = extractelement <2 x i16> %coords_lo, i32 0
810  %slice = extractelement <2 x i16> %coords_lo, i32 1
811  %mip = extractelement <2 x i16> %coords_hi, i32 0
812  call void @llvm.amdgcn.image.store.mip.1darray.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %slice, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
813  ret void
814}
815
816define amdgpu_ps void @store_mip_2darray(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords_lo, <2 x i16> %coords_hi) {
817; GFX9-LABEL: store_mip_2darray:
818; GFX9:       ; %bb.0: ; %main_body
819; GFX9-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x24,0xf0,0x04,0x00,0x00,0x00]
820; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
821;
822; GFX10-LABEL: store_mip_2darray:
823; GFX10:       ; %bb.0: ; %main_body
824; GFX10-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x28,0x1f,0x24,0xf0,0x04,0x00,0x00,0x40]
825; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
826;
827; GFX11-LABEL: store_mip_2darray:
828; GFX11:       ; %bb.0: ; %main_body
829; GFX11-NEXT:    image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x94,0x0f,0x1d,0xf0,0x04,0x00,0x00,0x00]
830; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
831;
832; GFX12-LABEL: store_mip_2darray:
833; GFX12:       ; %bb.0: ; %main_body
834; GFX12-NEXT:    image_store_mip v[0:3], [v4, v5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x45,0xc0,0xc1,0xd3,0x00,0x00,0x00,0x00,0x04,0x05,0x00,0x00]
835; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
836main_body:
837  %s = extractelement <2 x i16> %coords_lo, i32 0
838  %t = extractelement <2 x i16> %coords_lo, i32 1
839  %slice = extractelement <2 x i16> %coords_hi, i32 0
840  %mip = extractelement <2 x i16> %coords_hi, i32 1
841  call void @llvm.amdgcn.image.store.mip.2darray.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, i16 %t, i16 %slice, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
842  ret void
843}
844
845define amdgpu_ps <4 x float> @getresinfo_1d(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
846; GFX9-LABEL: getresinfo_1d:
847; GFX9:       ; %bb.0: ; %main_body
848; GFX9-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x38,0xf0,0x00,0x00,0x00,0x00]
849; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
850; GFX9-NEXT:    ; return to shader part epilog
851;
852; GFX10-LABEL: getresinfo_1d:
853; GFX10:       ; %bb.0: ; %main_body
854; GFX10-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x1f,0x38,0xf0,0x00,0x00,0x00,0x40]
855; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
856; GFX10-NEXT:    ; return to shader part epilog
857;
858; GFX11-LABEL: getresinfo_1d:
859; GFX11:       ; %bb.0: ; %main_body
860; GFX11-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x0f,0x5d,0xf0,0x00,0x00,0x00,0x00]
861; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
862; GFX11-NEXT:    ; return to shader part epilog
863;
864; GFX12-LABEL: getresinfo_1d:
865; GFX12:       ; %bb.0: ; %main_body
866; GFX12-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0xc0,0xc5,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
867; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
868; GFX12-NEXT:    ; return to shader part epilog
869main_body:
870  %mip = extractelement <2 x i16> %coords, i32 0
871  %v = call <4 x float> @llvm.amdgcn.image.getresinfo.1d.v4f32.i16(i32 15, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
872  ret <4 x float> %v
873}
874
875define amdgpu_ps <4 x float> @getresinfo_2d(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
876; GFX9-LABEL: getresinfo_2d:
877; GFX9:       ; %bb.0: ; %main_body
878; GFX9-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x38,0xf0,0x00,0x00,0x00,0x00]
879; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
880; GFX9-NEXT:    ; return to shader part epilog
881;
882; GFX10-LABEL: getresinfo_2d:
883; GFX10:       ; %bb.0: ; %main_body
884; GFX10-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x08,0x1f,0x38,0xf0,0x00,0x00,0x00,0x40]
885; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
886; GFX10-NEXT:    ; return to shader part epilog
887;
888; GFX11-LABEL: getresinfo_2d:
889; GFX11:       ; %bb.0: ; %main_body
890; GFX11-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm a16 ; encoding: [0x84,0x0f,0x5d,0xf0,0x00,0x00,0x00,0x00]
891; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
892; GFX11-NEXT:    ; return to shader part epilog
893;
894; GFX12-LABEL: getresinfo_2d:
895; GFX12:       ; %bb.0: ; %main_body
896; GFX12-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D a16 ; encoding: [0x41,0xc0,0xc5,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
897; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
898; GFX12-NEXT:    ; return to shader part epilog
899main_body:
900  %mip = extractelement <2 x i16> %coords, i32 0
901  %v = call <4 x float> @llvm.amdgcn.image.getresinfo.2d.v4f32.i16(i32 15, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
902  ret <4 x float> %v
903}
904
905define amdgpu_ps <4 x float> @getresinfo_3d(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
906; GFX9-LABEL: getresinfo_3d:
907; GFX9:       ; %bb.0: ; %main_body
908; GFX9-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x38,0xf0,0x00,0x00,0x00,0x00]
909; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
910; GFX9-NEXT:    ; return to shader part epilog
911;
912; GFX10-LABEL: getresinfo_3d:
913; GFX10:       ; %bb.0: ; %main_body
914; GFX10-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x10,0x1f,0x38,0xf0,0x00,0x00,0x00,0x40]
915; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
916; GFX10-NEXT:    ; return to shader part epilog
917;
918; GFX11-LABEL: getresinfo_3d:
919; GFX11:       ; %bb.0: ; %main_body
920; GFX11-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm a16 ; encoding: [0x88,0x0f,0x5d,0xf0,0x00,0x00,0x00,0x00]
921; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
922; GFX11-NEXT:    ; return to shader part epilog
923;
924; GFX12-LABEL: getresinfo_3d:
925; GFX12:       ; %bb.0: ; %main_body
926; GFX12-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D a16 ; encoding: [0x42,0xc0,0xc5,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
927; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
928; GFX12-NEXT:    ; return to shader part epilog
929main_body:
930  %mip = extractelement <2 x i16> %coords, i32 0
931  %v = call <4 x float> @llvm.amdgcn.image.getresinfo.3d.v4f32.i16(i32 15, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
932  ret <4 x float> %v
933}
934
935define amdgpu_ps <4 x float> @getresinfo_cube(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
936; GFX9-LABEL: getresinfo_cube:
937; GFX9:       ; %bb.0: ; %main_body
938; GFX9-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x38,0xf0,0x00,0x00,0x00,0x00]
939; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
940; GFX9-NEXT:    ; return to shader part epilog
941;
942; GFX10-LABEL: getresinfo_cube:
943; GFX10:       ; %bb.0: ; %main_body
944; GFX10-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x18,0x1f,0x38,0xf0,0x00,0x00,0x00,0x40]
945; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
946; GFX10-NEXT:    ; return to shader part epilog
947;
948; GFX11-LABEL: getresinfo_cube:
949; GFX11:       ; %bb.0: ; %main_body
950; GFX11-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm a16 ; encoding: [0x8c,0x0f,0x5d,0xf0,0x00,0x00,0x00,0x00]
951; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
952; GFX11-NEXT:    ; return to shader part epilog
953;
954; GFX12-LABEL: getresinfo_cube:
955; GFX12:       ; %bb.0: ; %main_body
956; GFX12-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE a16 ; encoding: [0x43,0xc0,0xc5,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
957; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
958; GFX12-NEXT:    ; return to shader part epilog
959main_body:
960  %mip = extractelement <2 x i16> %coords, i32 0
961  %v = call <4 x float> @llvm.amdgcn.image.getresinfo.cube.v4f32.i16(i32 15, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
962  ret <4 x float> %v
963}
964
965define amdgpu_ps <4 x float> @getresinfo_1darray(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
966; GFX9-LABEL: getresinfo_1darray:
967; GFX9:       ; %bb.0: ; %main_body
968; GFX9-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x38,0xf0,0x00,0x00,0x00,0x00]
969; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
970; GFX9-NEXT:    ; return to shader part epilog
971;
972; GFX10-LABEL: getresinfo_1darray:
973; GFX10:       ; %bb.0: ; %main_body
974; GFX10-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x20,0x1f,0x38,0xf0,0x00,0x00,0x00,0x40]
975; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
976; GFX10-NEXT:    ; return to shader part epilog
977;
978; GFX11-LABEL: getresinfo_1darray:
979; GFX11:       ; %bb.0: ; %main_body
980; GFX11-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm a16 ; encoding: [0x90,0x0f,0x5d,0xf0,0x00,0x00,0x00,0x00]
981; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
982; GFX11-NEXT:    ; return to shader part epilog
983;
984; GFX12-LABEL: getresinfo_1darray:
985; GFX12:       ; %bb.0: ; %main_body
986; GFX12-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY a16 ; encoding: [0x44,0xc0,0xc5,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
987; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
988; GFX12-NEXT:    ; return to shader part epilog
989main_body:
990  %mip = extractelement <2 x i16> %coords, i32 0
991  %v = call <4 x float> @llvm.amdgcn.image.getresinfo.1darray.v4f32.i16(i32 15, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
992  ret <4 x float> %v
993}
994
995define amdgpu_ps <4 x float> @getresinfo_2darray(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
996; GFX9-LABEL: getresinfo_2darray:
997; GFX9:       ; %bb.0: ; %main_body
998; GFX9-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x38,0xf0,0x00,0x00,0x00,0x00]
999; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
1000; GFX9-NEXT:    ; return to shader part epilog
1001;
1002; GFX10-LABEL: getresinfo_2darray:
1003; GFX10:       ; %bb.0: ; %main_body
1004; GFX10-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x28,0x1f,0x38,0xf0,0x00,0x00,0x00,0x40]
1005; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
1006; GFX10-NEXT:    ; return to shader part epilog
1007;
1008; GFX11-LABEL: getresinfo_2darray:
1009; GFX11:       ; %bb.0: ; %main_body
1010; GFX11-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm a16 ; encoding: [0x94,0x0f,0x5d,0xf0,0x00,0x00,0x00,0x00]
1011; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
1012; GFX11-NEXT:    ; return to shader part epilog
1013;
1014; GFX12-LABEL: getresinfo_2darray:
1015; GFX12:       ; %bb.0: ; %main_body
1016; GFX12-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY a16 ; encoding: [0x45,0xc0,0xc5,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
1017; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
1018; GFX12-NEXT:    ; return to shader part epilog
1019main_body:
1020  %mip = extractelement <2 x i16> %coords, i32 0
1021  %v = call <4 x float> @llvm.amdgcn.image.getresinfo.2darray.v4f32.i16(i32 15, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
1022  ret <4 x float> %v
1023}
1024
1025define amdgpu_ps <4 x float> @getresinfo_2dmsaa(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
1026; GFX9-LABEL: getresinfo_2dmsaa:
1027; GFX9:       ; %bb.0: ; %main_body
1028; GFX9-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm a16 ; encoding: [0x00,0x9f,0x38,0xf0,0x00,0x00,0x00,0x00]
1029; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
1030; GFX9-NEXT:    ; return to shader part epilog
1031;
1032; GFX10-LABEL: getresinfo_2dmsaa:
1033; GFX10:       ; %bb.0: ; %main_body
1034; GFX10-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm a16 ; encoding: [0x30,0x1f,0x38,0xf0,0x00,0x00,0x00,0x40]
1035; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
1036; GFX10-NEXT:    ; return to shader part epilog
1037;
1038; GFX11-LABEL: getresinfo_2dmsaa:
1039; GFX11:       ; %bb.0: ; %main_body
1040; GFX11-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm a16 ; encoding: [0x98,0x0f,0x5d,0xf0,0x00,0x00,0x00,0x00]
1041; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
1042; GFX11-NEXT:    ; return to shader part epilog
1043;
1044; GFX12-LABEL: getresinfo_2dmsaa:
1045; GFX12:       ; %bb.0: ; %main_body
1046; GFX12-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA a16 ; encoding: [0x46,0xc0,0xc5,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
1047; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
1048; GFX12-NEXT:    ; return to shader part epilog
1049main_body:
1050  %mip = extractelement <2 x i16> %coords, i32 0
1051  %v = call <4 x float> @llvm.amdgcn.image.getresinfo.2dmsaa.v4f32.i16(i32 15, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
1052  ret <4 x float> %v
1053}
1054
1055define amdgpu_ps <4 x float> @getresinfo_2darraymsaa(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
1056; GFX9-LABEL: getresinfo_2darraymsaa:
1057; GFX9:       ; %bb.0: ; %main_body
1058; GFX9-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm a16 da ; encoding: [0x00,0xdf,0x38,0xf0,0x00,0x00,0x00,0x00]
1059; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
1060; GFX9-NEXT:    ; return to shader part epilog
1061;
1062; GFX10-LABEL: getresinfo_2darraymsaa:
1063; GFX10:       ; %bb.0: ; %main_body
1064; GFX10-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm a16 ; encoding: [0x38,0x1f,0x38,0xf0,0x00,0x00,0x00,0x40]
1065; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
1066; GFX10-NEXT:    ; return to shader part epilog
1067;
1068; GFX11-LABEL: getresinfo_2darraymsaa:
1069; GFX11:       ; %bb.0: ; %main_body
1070; GFX11-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm a16 ; encoding: [0x9c,0x0f,0x5d,0xf0,0x00,0x00,0x00,0x00]
1071; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
1072; GFX11-NEXT:    ; return to shader part epilog
1073;
1074; GFX12-LABEL: getresinfo_2darraymsaa:
1075; GFX12:       ; %bb.0: ; %main_body
1076; GFX12-NEXT:    image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY a16 ; encoding: [0x47,0xc0,0xc5,0xd3,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
1077; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
1078; GFX12-NEXT:    ; return to shader part epilog
1079main_body:
1080  %mip = extractelement <2 x i16> %coords, i32 0
1081  %v = call <4 x float> @llvm.amdgcn.image.getresinfo.2darraymsaa.v4f32.i16(i32 15, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
1082  ret <4 x float> %v
1083}
1084
1085define amdgpu_ps float @load_1d_V1(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
1086; GFX9-LABEL: load_1d_V1:
1087; GFX9:       ; %bb.0: ; %main_body
1088; GFX9-NEXT:    image_load v0, v0, s[0:7] dmask:0x8 unorm a16 ; encoding: [0x00,0x98,0x00,0xf0,0x00,0x00,0x00,0x00]
1089; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
1090; GFX9-NEXT:    ; return to shader part epilog
1091;
1092; GFX10-LABEL: load_1d_V1:
1093; GFX10:       ; %bb.0: ; %main_body
1094; GFX10-NEXT:    image_load v0, v0, s[0:7] dmask:0x8 dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x18,0x00,0xf0,0x00,0x00,0x00,0x40]
1095; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
1096; GFX10-NEXT:    ; return to shader part epilog
1097;
1098; GFX11-LABEL: load_1d_V1:
1099; GFX11:       ; %bb.0: ; %main_body
1100; GFX11-NEXT:    image_load v0, v0, s[0:7] dmask:0x8 dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x08,0x01,0xf0,0x00,0x00,0x00,0x00]
1101; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
1102; GFX11-NEXT:    ; return to shader part epilog
1103;
1104; GFX12-LABEL: load_1d_V1:
1105; GFX12:       ; %bb.0: ; %main_body
1106; GFX12-NEXT:    image_load v0, v0, s[0:7] dmask:0x8 dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0x00,0x00,0xd2,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
1107; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
1108; GFX12-NEXT:    ; return to shader part epilog
1109main_body:
1110  %s = extractelement <2 x i16> %coords, i32 0
1111  %v = call float @llvm.amdgcn.image.load.1d.f32.i16(i32 8, i16 %s, <8 x i32> %rsrc, i32 0, i32 0)
1112  ret float %v
1113}
1114
1115define amdgpu_ps <2 x float> @load_1d_V2(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
1116; GFX9-LABEL: load_1d_V2:
1117; GFX9:       ; %bb.0: ; %main_body
1118; GFX9-NEXT:    image_load v[0:1], v0, s[0:7] dmask:0x9 unorm a16 ; encoding: [0x00,0x99,0x00,0xf0,0x00,0x00,0x00,0x00]
1119; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
1120; GFX9-NEXT:    ; return to shader part epilog
1121;
1122; GFX10-LABEL: load_1d_V2:
1123; GFX10:       ; %bb.0: ; %main_body
1124; GFX10-NEXT:    image_load v[0:1], v0, s[0:7] dmask:0x9 dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x19,0x00,0xf0,0x00,0x00,0x00,0x40]
1125; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
1126; GFX10-NEXT:    ; return to shader part epilog
1127;
1128; GFX11-LABEL: load_1d_V2:
1129; GFX11:       ; %bb.0: ; %main_body
1130; GFX11-NEXT:    image_load v[0:1], v0, s[0:7] dmask:0x9 dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x09,0x01,0xf0,0x00,0x00,0x00,0x00]
1131; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
1132; GFX11-NEXT:    ; return to shader part epilog
1133;
1134; GFX12-LABEL: load_1d_V2:
1135; GFX12:       ; %bb.0: ; %main_body
1136; GFX12-NEXT:    image_load v[0:1], v0, s[0:7] dmask:0x9 dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0x00,0x40,0xd2,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
1137; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
1138; GFX12-NEXT:    ; return to shader part epilog
1139main_body:
1140  %s = extractelement <2 x i16> %coords, i32 0
1141  %v = call <2 x float> @llvm.amdgcn.image.load.1d.v2f32.i16(i32 9, i16 %s, <8 x i32> %rsrc, i32 0, i32 0)
1142  ret <2 x float> %v
1143}
1144
1145define amdgpu_ps void @store_1d_V1(<8 x i32> inreg %rsrc, float %vdata, <2 x i16> %coords) {
1146; GFX9-LABEL: store_1d_V1:
1147; GFX9:       ; %bb.0: ; %main_body
1148; GFX9-NEXT:    image_store v0, v1, s[0:7] dmask:0x2 unorm a16 ; encoding: [0x00,0x92,0x20,0xf0,0x01,0x00,0x00,0x00]
1149; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1150;
1151; GFX10-LABEL: store_1d_V1:
1152; GFX10:       ; %bb.0: ; %main_body
1153; GFX10-NEXT:    image_store v0, v1, s[0:7] dmask:0x2 dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x12,0x20,0xf0,0x01,0x00,0x00,0x40]
1154; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1155;
1156; GFX11-LABEL: store_1d_V1:
1157; GFX11:       ; %bb.0: ; %main_body
1158; GFX11-NEXT:    image_store v0, v1, s[0:7] dmask:0x2 dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x02,0x19,0xf0,0x01,0x00,0x00,0x00]
1159; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1160;
1161; GFX12-LABEL: store_1d_V1:
1162; GFX12:       ; %bb.0: ; %main_body
1163; GFX12-NEXT:    image_store v0, v1, s[0:7] dmask:0x2 dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0x80,0x81,0xd0,0x00,0x00,0x00,0x00,0x01,0x00,0x00,0x00]
1164; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1165main_body:
1166  %s = extractelement <2 x i16> %coords, i32 0
1167  call void @llvm.amdgcn.image.store.1d.f32.i16(float %vdata, i32 2, i16 %s, <8 x i32> %rsrc, i32 0, i32 0)
1168  ret void
1169}
1170
1171define amdgpu_ps void @store_1d_V2(<8 x i32> inreg %rsrc, <2 x float> %vdata, <2 x i16> %coords) {
1172; GFX9-LABEL: store_1d_V2:
1173; GFX9:       ; %bb.0: ; %main_body
1174; GFX9-NEXT:    image_store v[0:1], v2, s[0:7] dmask:0xc unorm a16 ; encoding: [0x00,0x9c,0x20,0xf0,0x02,0x00,0x00,0x00]
1175; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1176;
1177; GFX10-LABEL: store_1d_V2:
1178; GFX10:       ; %bb.0: ; %main_body
1179; GFX10-NEXT:    image_store v[0:1], v2, s[0:7] dmask:0xc dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x00,0x1c,0x20,0xf0,0x02,0x00,0x00,0x40]
1180; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1181;
1182; GFX11-LABEL: store_1d_V2:
1183; GFX11:       ; %bb.0: ; %main_body
1184; GFX11-NEXT:    image_store v[0:1], v2, s[0:7] dmask:0xc dim:SQ_RSRC_IMG_1D unorm a16 ; encoding: [0x80,0x0c,0x19,0xf0,0x02,0x00,0x00,0x00]
1185; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1186;
1187; GFX12-LABEL: store_1d_V2:
1188; GFX12:       ; %bb.0: ; %main_body
1189; GFX12-NEXT:    image_store v[0:1], v2, s[0:7] dmask:0xc dim:SQ_RSRC_IMG_1D a16 ; encoding: [0x40,0x80,0x01,0xd3,0x00,0x00,0x00,0x00,0x02,0x00,0x00,0x00]
1190; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1191main_body:
1192  %s = extractelement <2 x i16> %coords, i32 0
1193  call void @llvm.amdgcn.image.store.1d.v2f32.i16(<2 x float> %vdata, i32 12, i16 %s, <8 x i32> %rsrc, i32 0, i32 0)
1194  ret void
1195}
1196
1197define amdgpu_ps <4 x float> @load_1d_glc(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
1198; GFX9-LABEL: load_1d_glc:
1199; GFX9:       ; %bb.0: ; %main_body
1200; GFX9-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc a16 ; encoding: [0x00,0xbf,0x00,0xf0,0x00,0x00,0x00,0x00]
1201; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
1202; GFX9-NEXT:    ; return to shader part epilog
1203;
1204; GFX10-LABEL: load_1d_glc:
1205; GFX10:       ; %bb.0: ; %main_body
1206; GFX10-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc a16 ; encoding: [0x00,0x3f,0x00,0xf0,0x00,0x00,0x00,0x40]
1207; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
1208; GFX10-NEXT:    ; return to shader part epilog
1209;
1210; GFX11-LABEL: load_1d_glc:
1211; GFX11:       ; %bb.0: ; %main_body
1212; GFX11-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc a16 ; encoding: [0x80,0x4f,0x01,0xf0,0x00,0x00,0x00,0x00]
1213; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
1214; GFX11-NEXT:    ; return to shader part epilog
1215;
1216; GFX12-LABEL: load_1d_glc:
1217; GFX12:       ; %bb.0: ; %main_body
1218; GFX12-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D th:TH_LOAD_NT a16 ; encoding: [0x40,0x00,0xc0,0xd3,0x00,0x00,0x10,0x00,0x00,0x00,0x00,0x00]
1219; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
1220; GFX12-NEXT:    ; return to shader part epilog
1221main_body:
1222  %s = extractelement <2 x i16> %coords, i32 0
1223  %v = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i16(i32 15, i16 %s, <8 x i32> %rsrc, i32 0, i32 1)
1224  ret <4 x float> %v
1225}
1226
1227define amdgpu_ps <4 x float> @load_1d_slc(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
1228; GFX9-LABEL: load_1d_slc:
1229; GFX9:       ; %bb.0: ; %main_body
1230; GFX9-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf unorm slc a16 ; encoding: [0x00,0x9f,0x00,0xf2,0x00,0x00,0x00,0x00]
1231; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
1232; GFX9-NEXT:    ; return to shader part epilog
1233;
1234; GFX10-LABEL: load_1d_slc:
1235; GFX10:       ; %bb.0: ; %main_body
1236; GFX10-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm slc a16 ; encoding: [0x00,0x1f,0x00,0xf2,0x00,0x00,0x00,0x40]
1237; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
1238; GFX10-NEXT:    ; return to shader part epilog
1239;
1240; GFX11-LABEL: load_1d_slc:
1241; GFX11:       ; %bb.0: ; %main_body
1242; GFX11-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm slc a16 ; encoding: [0x80,0x1f,0x01,0xf0,0x00,0x00,0x00,0x00]
1243; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
1244; GFX11-NEXT:    ; return to shader part epilog
1245;
1246; GFX12-LABEL: load_1d_slc:
1247; GFX12:       ; %bb.0: ; %main_body
1248; GFX12-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D th:TH_LOAD_HT a16 ; encoding: [0x40,0x00,0xc0,0xd3,0x00,0x00,0x20,0x00,0x00,0x00,0x00,0x00]
1249; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
1250; GFX12-NEXT:    ; return to shader part epilog
1251main_body:
1252  %s = extractelement <2 x i16> %coords, i32 0
1253  %v = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i16(i32 15, i16 %s, <8 x i32> %rsrc, i32 0, i32 2)
1254  ret <4 x float> %v
1255}
1256
1257define amdgpu_ps <4 x float> @load_1d_glc_slc(<8 x i32> inreg %rsrc, <2 x i16> %coords) {
1258; GFX9-LABEL: load_1d_glc_slc:
1259; GFX9:       ; %bb.0: ; %main_body
1260; GFX9-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc slc a16 ; encoding: [0x00,0xbf,0x00,0xf2,0x00,0x00,0x00,0x00]
1261; GFX9-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x0f,0x8c,0xbf]
1262; GFX9-NEXT:    ; return to shader part epilog
1263;
1264; GFX10-LABEL: load_1d_glc_slc:
1265; GFX10:       ; %bb.0: ; %main_body
1266; GFX10-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc slc a16 ; encoding: [0x00,0x3f,0x00,0xf2,0x00,0x00,0x00,0x40]
1267; GFX10-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
1268; GFX10-NEXT:    ; return to shader part epilog
1269;
1270; GFX11-LABEL: load_1d_glc_slc:
1271; GFX11:       ; %bb.0: ; %main_body
1272; GFX11-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc slc a16 ; encoding: [0x80,0x5f,0x01,0xf0,0x00,0x00,0x00,0x00]
1273; GFX11-NEXT:    s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]
1274; GFX11-NEXT:    ; return to shader part epilog
1275;
1276; GFX12-LABEL: load_1d_glc_slc:
1277; GFX12:       ; %bb.0: ; %main_body
1278; GFX12-NEXT:    image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D th:TH_LOAD_LU a16 ; encoding: [0x40,0x00,0xc0,0xd3,0x00,0x00,0x30,0x00,0x00,0x00,0x00,0x00]
1279; GFX12-NEXT:    s_wait_loadcnt 0x0 ; encoding: [0x00,0x00,0xc0,0xbf]
1280; GFX12-NEXT:    ; return to shader part epilog
1281main_body:
1282  %s = extractelement <2 x i16> %coords, i32 0
1283  %v = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i16(i32 15, i16 %s, <8 x i32> %rsrc, i32 0, i32 3)
1284  ret <4 x float> %v
1285}
1286
1287define amdgpu_ps void @store_1d_glc(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords) {
1288; GFX9-LABEL: store_1d_glc:
1289; GFX9:       ; %bb.0: ; %main_body
1290; GFX9-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc a16 ; encoding: [0x00,0xbf,0x20,0xf0,0x04,0x00,0x00,0x00]
1291; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1292;
1293; GFX10-LABEL: store_1d_glc:
1294; GFX10:       ; %bb.0: ; %main_body
1295; GFX10-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc a16 ; encoding: [0x00,0x3f,0x20,0xf0,0x04,0x00,0x00,0x40]
1296; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1297;
1298; GFX11-LABEL: store_1d_glc:
1299; GFX11:       ; %bb.0: ; %main_body
1300; GFX11-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc a16 ; encoding: [0x80,0x4f,0x19,0xf0,0x04,0x00,0x00,0x00]
1301; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1302;
1303; GFX12-LABEL: store_1d_glc:
1304; GFX12:       ; %bb.0: ; %main_body
1305; GFX12-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D th:TH_STORE_NT a16 ; encoding: [0x40,0x80,0xc1,0xd3,0x00,0x00,0x10,0x00,0x04,0x00,0x00,0x00]
1306; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1307main_body:
1308  %s = extractelement <2 x i16> %coords, i32 0
1309  call void @llvm.amdgcn.image.store.1d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, <8 x i32> %rsrc, i32 0, i32 1)
1310  ret void
1311}
1312
1313define amdgpu_ps void @store_1d_slc(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords) {
1314; GFX9-LABEL: store_1d_slc:
1315; GFX9:       ; %bb.0: ; %main_body
1316; GFX9-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf unorm slc a16 ; encoding: [0x00,0x9f,0x20,0xf2,0x04,0x00,0x00,0x00]
1317; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1318;
1319; GFX10-LABEL: store_1d_slc:
1320; GFX10:       ; %bb.0: ; %main_body
1321; GFX10-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm slc a16 ; encoding: [0x00,0x1f,0x20,0xf2,0x04,0x00,0x00,0x40]
1322; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1323;
1324; GFX11-LABEL: store_1d_slc:
1325; GFX11:       ; %bb.0: ; %main_body
1326; GFX11-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm slc a16 ; encoding: [0x80,0x1f,0x19,0xf0,0x04,0x00,0x00,0x00]
1327; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1328;
1329; GFX12-LABEL: store_1d_slc:
1330; GFX12:       ; %bb.0: ; %main_body
1331; GFX12-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D th:TH_STORE_HT a16 ; encoding: [0x40,0x80,0xc1,0xd3,0x00,0x00,0x20,0x00,0x04,0x00,0x00,0x00]
1332; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1333main_body:
1334  %s = extractelement <2 x i16> %coords, i32 0
1335  call void @llvm.amdgcn.image.store.1d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, <8 x i32> %rsrc, i32 0, i32 2)
1336  ret void
1337}
1338
1339define amdgpu_ps void @store_1d_glc_slc(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords) {
1340; GFX9-LABEL: store_1d_glc_slc:
1341; GFX9:       ; %bb.0: ; %main_body
1342; GFX9-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc slc a16 ; encoding: [0x00,0xbf,0x20,0xf2,0x04,0x00,0x00,0x00]
1343; GFX9-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1344;
1345; GFX10-LABEL: store_1d_glc_slc:
1346; GFX10:       ; %bb.0: ; %main_body
1347; GFX10-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc slc a16 ; encoding: [0x00,0x3f,0x20,0xf2,0x04,0x00,0x00,0x40]
1348; GFX10-NEXT:    s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
1349;
1350; GFX11-LABEL: store_1d_glc_slc:
1351; GFX11:       ; %bb.0: ; %main_body
1352; GFX11-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc slc a16 ; encoding: [0x80,0x5f,0x19,0xf0,0x04,0x00,0x00,0x00]
1353; GFX11-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1354;
1355; GFX12-LABEL: store_1d_glc_slc:
1356; GFX12:       ; %bb.0: ; %main_body
1357; GFX12-NEXT:    image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D th:TH_STORE_RT_WB a16 ; encoding: [0x40,0x80,0xc1,0xd3,0x00,0x00,0x30,0x00,0x04,0x00,0x00,0x00]
1358; GFX12-NEXT:    s_endpgm ; encoding: [0x00,0x00,0xb0,0xbf]
1359main_body:
1360  %s = extractelement <2 x i16> %coords, i32 0
1361  call void @llvm.amdgcn.image.store.1d.v4f32.i16(<4 x float> %vdata, i32 15, i16 %s, <8 x i32> %rsrc, i32 0, i32 3)
1362  ret void
1363}
1364
1365define amdgpu_ps <4 x float> @getresinfo_dmask0(<8 x i32> inreg %rsrc, <4 x float> %vdata, <2 x i16> %coords) #0 {
1366; GFX9-LABEL: getresinfo_dmask0:
1367; GFX9:       ; %bb.0: ; %main_body
1368; GFX9-NEXT:    ; return to shader part epilog
1369;
1370; GFX10-LABEL: getresinfo_dmask0:
1371; GFX10:       ; %bb.0: ; %main_body
1372; GFX10-NEXT:    ; return to shader part epilog
1373;
1374; GFX11-LABEL: getresinfo_dmask0:
1375; GFX11:       ; %bb.0: ; %main_body
1376; GFX11-NEXT:    ; return to shader part epilog
1377;
1378; GFX12-LABEL: getresinfo_dmask0:
1379; GFX12:       ; %bb.0: ; %main_body
1380; GFX12-NEXT:    ; return to shader part epilog
1381main_body:
1382  %mip = extractelement <2 x i16> %coords, i32 0
1383  %r = call <4 x float> @llvm.amdgcn.image.getresinfo.1d.v4f32.i16(i32 0, i16 %mip, <8 x i32> %rsrc, i32 0, i32 0)
1384  ret <4 x float> %r
1385}
1386
1387declare <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #1
1388declare <4 x float> @llvm.amdgcn.image.load.2d.v4f32.i16(i32, i16, i16, <8 x i32>, i32, i32) #1
1389declare <4 x float> @llvm.amdgcn.image.load.3d.v4f32.i16(i32, i16, i16, i16, <8 x i32>, i32, i32) #1
1390declare <4 x float> @llvm.amdgcn.image.load.cube.v4f32.i16(i32, i16, i16, i16, <8 x i32>, i32, i32) #1
1391declare <4 x float> @llvm.amdgcn.image.load.1darray.v4f32.i16(i32, i16, i16, <8 x i32>, i32, i32) #1
1392declare <4 x float> @llvm.amdgcn.image.load.2darray.v4f32.i16(i32, i16, i16, i16, <8 x i32>, i32, i32) #1
1393declare <4 x float> @llvm.amdgcn.image.load.2dmsaa.v4f32.i16(i32, i16, i16, i16, <8 x i32>, i32, i32) #1
1394declare <4 x float> @llvm.amdgcn.image.load.2darraymsaa.v4f32.i16(i32, i16, i16, i16, i16, <8 x i32>, i32, i32) #1
1395
1396declare <4 x float> @llvm.amdgcn.image.load.mip.1d.v4f32.i16(i32, i16, i16, <8 x i32>, i32, i32) #1
1397declare <4 x float> @llvm.amdgcn.image.load.mip.2d.v4f32.i16(i32, i16, i16, i16, <8 x i32>, i32, i32) #1
1398declare <4 x float> @llvm.amdgcn.image.load.mip.3d.v4f32.i16(i32, i16, i16, i16, i16, <8 x i32>, i32, i32) #1
1399declare <4 x float> @llvm.amdgcn.image.load.mip.cube.v4f32.i16(i32, i16, i16, i16, i16, <8 x i32>, i32, i32) #1
1400declare <4 x float> @llvm.amdgcn.image.load.mip.1darray.v4f32.i16(i32, i16, i16, i16, <8 x i32>, i32, i32) #1
1401declare <4 x float> @llvm.amdgcn.image.load.mip.2darray.v4f32.i16(i32, i16, i16, i16, i16, <8 x i32>, i32, i32) #1
1402
1403declare void @llvm.amdgcn.image.store.1d.v4f32.i16(<4 x float>, i32, i16, <8 x i32>, i32, i32) #0
1404declare void @llvm.amdgcn.image.store.2d.v4f32.i16(<4 x float>, i32, i16, i16, <8 x i32>, i32, i32) #0
1405declare void @llvm.amdgcn.image.store.3d.v4f32.i16(<4 x float>, i32, i16, i16, i16, <8 x i32>, i32, i32) #0
1406declare void @llvm.amdgcn.image.store.cube.v4f32.i16(<4 x float>, i32, i16, i16, i16, <8 x i32>, i32, i32) #0
1407declare void @llvm.amdgcn.image.store.1darray.v4f32.i16(<4 x float>, i32, i16, i16, <8 x i32>, i32, i32) #0
1408declare void @llvm.amdgcn.image.store.2darray.v4f32.i16(<4 x float>, i32, i16, i16, i16, <8 x i32>, i32, i32) #0
1409declare void @llvm.amdgcn.image.store.2dmsaa.v4f32.i16(<4 x float>, i32, i16, i16, i16, <8 x i32>, i32, i32) #0
1410declare void @llvm.amdgcn.image.store.2darraymsaa.v4f32.i16(<4 x float>, i32, i16, i16, i16, i16, <8 x i32>, i32, i32) #0
1411
1412declare void @llvm.amdgcn.image.store.mip.1d.v4f32.i16(<4 x float>, i32, i16, i16, <8 x i32>, i32, i32) #0
1413declare void @llvm.amdgcn.image.store.mip.2d.v4f32.i16(<4 x float>, i32, i16, i16, i16, <8 x i32>, i32, i32) #0
1414declare void @llvm.amdgcn.image.store.mip.3d.v4f32.i16(<4 x float>, i32, i16, i16, i16, i16, <8 x i32>, i32, i32) #0
1415declare void @llvm.amdgcn.image.store.mip.cube.v4f32.i16(<4 x float>, i32, i16, i16, i16, i16, <8 x i32>, i32, i32) #0
1416declare void @llvm.amdgcn.image.store.mip.1darray.v4f32.i16(<4 x float>, i32, i16, i16, i16, <8 x i32>, i32, i32) #0
1417declare void @llvm.amdgcn.image.store.mip.2darray.v4f32.i16(<4 x float>, i32, i16, i16, i16, i16, <8 x i32>, i32, i32) #0
1418
1419declare <4 x float> @llvm.amdgcn.image.getresinfo.1d.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #2
1420declare <4 x float> @llvm.amdgcn.image.getresinfo.2d.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #2
1421declare <4 x float> @llvm.amdgcn.image.getresinfo.3d.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #2
1422declare <4 x float> @llvm.amdgcn.image.getresinfo.cube.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #2
1423declare <4 x float> @llvm.amdgcn.image.getresinfo.1darray.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #2
1424declare <4 x float> @llvm.amdgcn.image.getresinfo.2darray.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #2
1425declare <4 x float> @llvm.amdgcn.image.getresinfo.2dmsaa.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #2
1426declare <4 x float> @llvm.amdgcn.image.getresinfo.2darraymsaa.v4f32.i16(i32, i16, <8 x i32>, i32, i32) #2
1427
1428declare float @llvm.amdgcn.image.load.1d.f32.i16(i32, i16, <8 x i32>, i32, i32) #1
1429declare float @llvm.amdgcn.image.load.2d.f32.i16(i32, i16, i16, <8 x i32>, i32, i32) #1
1430declare <2 x float> @llvm.amdgcn.image.load.1d.v2f32.i16(i32, i16, <8 x i32>, i32, i32) #1
1431declare void @llvm.amdgcn.image.store.1d.f32.i16(float, i32, i16, <8 x i32>, i32, i32) #0
1432declare void @llvm.amdgcn.image.store.1d.v2f32.i16(<2 x float>, i32, i16, <8 x i32>, i32, i32) #0
1433
1434attributes #0 = { nounwind }
1435attributes #1 = { nounwind readonly }
1436attributes #2 = { nounwind readnone }
1437