1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc -global-isel -mtriple=amdgcn -mcpu=gfx1010 -verify-machineinstrs < %s | FileCheck -check-prefix=GFX10 %s
4 define amdgpu_ps <4 x float> @sample_cd_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %dsdh, half %dsdv, float %s) {
5 ; GFX10-LABEL: sample_cd_1d:
6 ; GFX10: ; %bb.0: ; %main_body
7 ; GFX10-NEXT: image_sample_cd_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
8 ; GFX10-NEXT: s_waitcnt vmcnt(0)
9 ; GFX10-NEXT: ; return to shader part epilog
11 %v = call <4 x float> @llvm.amdgcn.image.sample.cd.1d.v4f32.f16.f32(i32 15, half %dsdh, half %dsdv, float %s, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)
15 define amdgpu_ps <4 x float> @sample_cd_2d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t) {
16 ; GFX10-LABEL: sample_cd_2d:
17 ; GFX10: ; %bb.0: ; %main_body
18 ; GFX10-NEXT: v_and_b32_e32 v0, 0xffff, v0
19 ; GFX10-NEXT: v_and_b32_e32 v2, 0xffff, v2
20 ; GFX10-NEXT: v_lshl_or_b32 v0, v1, 16, v0
21 ; GFX10-NEXT: v_lshl_or_b32 v1, v3, 16, v2
22 ; GFX10-NEXT: image_sample_cd_g16 v[0:3], [v0, v1, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
23 ; GFX10-NEXT: s_waitcnt vmcnt(0)
24 ; GFX10-NEXT: ; return to shader part epilog
26 %v = call <4 x float> @llvm.amdgcn.image.sample.cd.2d.v4f32.f16.f32(i32 15, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)
30 define amdgpu_ps <4 x float> @sample_c_cd_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, float %zcompare, half %dsdh, half %dsdv, float %s) {
31 ; GFX10-LABEL: sample_c_cd_1d:
32 ; GFX10: ; %bb.0: ; %main_body
33 ; GFX10-NEXT: image_sample_c_cd_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
34 ; GFX10-NEXT: s_waitcnt vmcnt(0)
35 ; GFX10-NEXT: ; return to shader part epilog
37 %v = call <4 x float> @llvm.amdgcn.image.sample.c.cd.1d.v4f32.f16.f32(i32 15, float %zcompare, half %dsdh, half %dsdv, float %s, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)
41 define amdgpu_ps <4 x float> @sample_c_cd_2d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, float %zcompare, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t) {
42 ; GFX10-LABEL: sample_c_cd_2d:
43 ; GFX10: ; %bb.0: ; %main_body
44 ; GFX10-NEXT: v_and_b32_e32 v1, 0xffff, v1
45 ; GFX10-NEXT: v_and_b32_e32 v3, 0xffff, v3
46 ; GFX10-NEXT: v_lshl_or_b32 v1, v2, 16, v1
47 ; GFX10-NEXT: v_lshl_or_b32 v2, v4, 16, v3
48 ; GFX10-NEXT: image_sample_c_cd_g16 v[0:3], [v0, v1, v2, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
49 ; GFX10-NEXT: s_waitcnt vmcnt(0)
50 ; GFX10-NEXT: ; return to shader part epilog
52 %v = call <4 x float> @llvm.amdgcn.image.sample.c.cd.2d.v4f32.f16.f32(i32 15, float %zcompare, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)
56 define amdgpu_ps <4 x float> @sample_cd_cl_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %dsdh, half %dsdv, float %s, float %clamp) {
57 ; GFX10-LABEL: sample_cd_cl_1d:
58 ; GFX10: ; %bb.0: ; %main_body
59 ; GFX10-NEXT: image_sample_cd_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
60 ; GFX10-NEXT: s_waitcnt vmcnt(0)
61 ; GFX10-NEXT: ; return to shader part epilog
63 %v = call <4 x float> @llvm.amdgcn.image.sample.cd.cl.1d.v4f32.f16.f32(i32 15, half %dsdh, half %dsdv, float %s, float %clamp, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)
67 define amdgpu_ps <4 x float> @sample_cd_cl_2d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, float %clamp) {
68 ; GFX10-LABEL: sample_cd_cl_2d:
69 ; GFX10: ; %bb.0: ; %main_body
70 ; GFX10-NEXT: v_and_b32_e32 v0, 0xffff, v0
71 ; GFX10-NEXT: v_and_b32_e32 v2, 0xffff, v2
72 ; GFX10-NEXT: v_lshl_or_b32 v0, v1, 16, v0
73 ; GFX10-NEXT: v_lshl_or_b32 v1, v3, 16, v2
74 ; GFX10-NEXT: image_sample_cd_cl_g16 v[0:3], [v0, v1, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
75 ; GFX10-NEXT: s_waitcnt vmcnt(0)
76 ; GFX10-NEXT: ; return to shader part epilog
78 %v = call <4 x float> @llvm.amdgcn.image.sample.cd.cl.2d.v4f32.f16.f32(i32 15, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, float %clamp, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)
82 define amdgpu_ps <4 x float> @sample_c_cd_cl_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, float %zcompare, half %dsdh, half %dsdv, float %s, float %clamp) {
83 ; GFX10-LABEL: sample_c_cd_cl_1d:
84 ; GFX10: ; %bb.0: ; %main_body
85 ; GFX10-NEXT: image_sample_c_cd_cl_g16 v[0:3], v[0:4], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D
86 ; GFX10-NEXT: s_waitcnt vmcnt(0)
87 ; GFX10-NEXT: ; return to shader part epilog
89 %v = call <4 x float> @llvm.amdgcn.image.sample.c.cd.cl.1d.v4f32.f16.f32(i32 15, float %zcompare, half %dsdh, half %dsdv, float %s, float %clamp, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)
93 define amdgpu_ps <4 x float> @sample_c_cd_cl_2d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, float %zcompare, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, float %clamp) {
94 ; GFX10-LABEL: sample_c_cd_cl_2d:
95 ; GFX10: ; %bb.0: ; %main_body
96 ; GFX10-NEXT: v_mov_b32_e32 v8, v2
97 ; GFX10-NEXT: v_mov_b32_e32 v2, v0
98 ; GFX10-NEXT: v_and_b32_e32 v0, 0xffff, v1
99 ; GFX10-NEXT: v_and_b32_e32 v1, 0xffff, v3
100 ; GFX10-NEXT: v_lshl_or_b32 v3, v8, 16, v0
101 ; GFX10-NEXT: v_lshl_or_b32 v4, v4, 16, v1
102 ; GFX10-NEXT: image_sample_c_cd_cl_g16 v[0:3], v[2:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D
103 ; GFX10-NEXT: s_waitcnt vmcnt(0)
104 ; GFX10-NEXT: ; return to shader part epilog
106 %v = call <4 x float> @llvm.amdgcn.image.sample.c.cd.cl.2d.v4f32.f16.f32(i32 15, float %zcompare, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, float %clamp, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)
110 declare <4 x float> @llvm.amdgcn.image.sample.cd.1d.v4f32.f16.f32(i32, half, half, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1
111 declare <4 x float> @llvm.amdgcn.image.sample.cd.2d.v4f32.f16.f32(i32, half, half, half, half, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1
112 declare <4 x float> @llvm.amdgcn.image.sample.c.cd.1d.v4f32.f16.f32(i32, float, half, half, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1
113 declare <4 x float> @llvm.amdgcn.image.sample.c.cd.2d.v4f32.f16.f32(i32, float, half, half, half, half, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1
114 declare <4 x float> @llvm.amdgcn.image.sample.cd.cl.1d.v4f32.f16.f32(i32, half, half, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1
115 declare <4 x float> @llvm.amdgcn.image.sample.cd.cl.2d.v4f32.f16.f32(i32, half, half, half, half, float, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1
116 declare <4 x float> @llvm.amdgcn.image.sample.c.cd.cl.1d.v4f32.f16.f32(i32, float, half, half, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1
117 declare <4 x float> @llvm.amdgcn.image.sample.c.cd.cl.2d.v4f32.f16.f32(i32, float, half, half, half, half, float, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1
119 attributes #0 = { nounwind }
120 attributes #1 = { nounwind readonly }
121 attributes #2 = { nounwind readnone }