449 lines · plain
1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py2; RUN: llc -mtriple=amdgcn -mcpu=gfx1010 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX10 %s3; RUN: llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=+real-true16 -amdgpu-enable-vopd=0 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX11,GFX11-TRUE16 %s4; RUN: llc -mtriple=amdgcn -mcpu=gfx1100 -mattr=-real-true16 -amdgpu-enable-vopd=0 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX11,GFX11-FAKE16 %s5; RUN: llc -mtriple=amdgcn -mcpu=gfx1200 -mattr=+real-true16 -amdgpu-enable-vopd=0 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX12,GFX12-TRUE16 %s6; RUN: llc -mtriple=amdgcn -mcpu=gfx1200 -mattr=-real-true16 -amdgpu-enable-vopd=0 -show-mc-encoding < %s | FileCheck -check-prefixes=GFX12,GFX12-FAKE16 %s7 8define amdgpu_ps <4 x float> @sample_d_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %dsdh, half %dsdv, float %s) {9; GFX10-LABEL: sample_d_1d:10; GFX10: ; %bb.0: ; %main_body11; GFX10-NEXT: image_sample_d_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00]12; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]13; GFX10-NEXT: ; return to shader part epilog14;15; GFX11-LABEL: sample_d_1d:16; GFX11: ; %bb.0: ; %main_body17; GFX11-NEXT: image_sample_d_g16 v[0:3], v[0:2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x0f,0xe4,0xf0,0x00,0x00,0x00,0x08]18; GFX11-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]19; GFX11-NEXT: ; return to shader part epilog20;21; GFX12-LABEL: sample_d_1d:22; GFX12: ; %bb.0: ; %main_body23; GFX12-NEXT: image_sample_d_g16 v[0:3], [v0, v1, v2], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x40,0xce,0xe7,0x00,0x00,0x00,0x04,0x00,0x01,0x02,0x00]24; GFX12-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]25; GFX12-NEXT: ; return to shader part epilog26main_body:27 %v = call <4 x float> @llvm.amdgcn.image.sample.d.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)28 ret <4 x float> %v29}30 31define amdgpu_ps <4 x float> @sample_d_2d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t) {32; GFX10-LABEL: sample_d_2d:33; GFX10: ; %bb.0: ; %main_body34; GFX10-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd7,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]35; GFX10-NEXT: v_perm_b32 v0, v1, v0, 0x5040100 ; encoding: [0x00,0x00,0x44,0xd7,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]36; GFX10-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0x88,0xf0,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x00]37; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]38; GFX10-NEXT: ; return to shader part epilog39;40; GFX11-TRUE16-LABEL: sample_d_2d:41; GFX11-TRUE16: ; %bb.0: ; %main_body42; GFX11-TRUE16-NEXT: v_mov_b16_e32 v2.h, v3.l ; encoding: [0x03,0x39,0x04,0x7f]43; GFX11-TRUE16-NEXT: v_mov_b16_e32 v0.h, v1.l ; encoding: [0x01,0x39,0x00,0x7f]44; GFX11-TRUE16-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x05,0x0f,0xe4,0xf0,0x00,0x00,0x00,0x08,0x02,0x04,0x05,0x00]45; GFX11-TRUE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]46; GFX11-TRUE16-NEXT: ; return to shader part epilog47;48; GFX11-FAKE16-LABEL: sample_d_2d:49; GFX11-FAKE16: ; %bb.0: ; %main_body50; GFX11-FAKE16-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd6,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]51; GFX11-FAKE16-NEXT: v_perm_b32 v0, v1, v0, 0x5040100 ; encoding: [0x00,0x00,0x44,0xd6,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]52; GFX11-FAKE16-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x05,0x0f,0xe4,0xf0,0x00,0x00,0x00,0x08,0x02,0x04,0x05,0x00]53; GFX11-FAKE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]54; GFX11-FAKE16-NEXT: ; return to shader part epilog55;56; GFX12-TRUE16-LABEL: sample_d_2d:57; GFX12-TRUE16: ; %bb.0: ; %main_body58; GFX12-TRUE16-NEXT: v_mov_b16_e32 v2.h, v3.l ; encoding: [0x03,0x39,0x04,0x7f]59; GFX12-TRUE16-NEXT: v_mov_b16_e32 v0.h, v1.l ; encoding: [0x01,0x39,0x00,0x7f]60; GFX12-TRUE16-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x01,0x40,0xce,0xe7,0x00,0x00,0x00,0x04,0x00,0x02,0x04,0x05]61; GFX12-TRUE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]62; GFX12-TRUE16-NEXT: ; return to shader part epilog63;64; GFX12-FAKE16-LABEL: sample_d_2d:65; GFX12-FAKE16: ; %bb.0: ; %main_body66; GFX12-FAKE16-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd6,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]67; GFX12-FAKE16-NEXT: v_perm_b32 v0, v1, v0, 0x5040100 ; encoding: [0x00,0x00,0x44,0xd6,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]68; GFX12-FAKE16-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v4, v5], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x01,0x40,0xce,0xe7,0x00,0x00,0x00,0x04,0x00,0x02,0x04,0x05]69; GFX12-FAKE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]70; GFX12-FAKE16-NEXT: ; return to shader part epilog71main_body:72 %v = call <4 x float> @llvm.amdgcn.image.sample.d.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)73 ret <4 x float> %v74}75 76define amdgpu_ps <4 x float> @sample_d_3d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %dsdh, half %dtdh, half %drdh, half %dsdv, half %dtdv, half %drdv, float %s, float %t, float %r) {77; GFX10-LABEL: sample_d_3d:78; GFX10: ; %bb.0: ; %main_body79; GFX10-NEXT: v_mov_b32_e32 v9, v3 ; encoding: [0x03,0x03,0x12,0x7e]80; GFX10-NEXT: v_mov_b32_e32 v3, v2 ; encoding: [0x02,0x03,0x06,0x7e]81; GFX10-NEXT: v_perm_b32 v2, v1, v0, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd7,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]82; GFX10-NEXT: v_perm_b32 v4, v4, v9, 0x5040100 ; encoding: [0x04,0x00,0x44,0xd7,0x04,0x13,0xfe,0x03,0x00,0x01,0x04,0x05]83; GFX10-NEXT: image_sample_d_g16 v[0:3], v[2:8], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D ; encoding: [0x11,0x0f,0x88,0xf0,0x02,0x00,0x40,0x00]84; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]85; GFX10-NEXT: ; return to shader part epilog86;87; GFX11-TRUE16-LABEL: sample_d_3d:88; GFX11-TRUE16: ; %bb.0: ; %main_body89; GFX11-TRUE16-NEXT: v_mov_b16_e32 v3.h, v4.l ; encoding: [0x04,0x39,0x06,0x7f]90; GFX11-TRUE16-NEXT: v_mov_b16_e32 v0.h, v1.l ; encoding: [0x01,0x39,0x00,0x7f]91; GFX11-TRUE16-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v3, v5, v[6:8]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D ; encoding: [0x09,0x0f,0xe4,0xf0,0x00,0x00,0x00,0x08,0x02,0x03,0x05,0x06]92; GFX11-TRUE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]93; GFX11-TRUE16-NEXT: ; return to shader part epilog94;95; GFX11-FAKE16-LABEL: sample_d_3d:96; GFX11-FAKE16: ; %bb.0: ; %main_body97; GFX11-FAKE16-NEXT: v_perm_b32 v3, v4, v3, 0x5040100 ; encoding: [0x03,0x00,0x44,0xd6,0x04,0x07,0xfe,0x03,0x00,0x01,0x04,0x05]98; GFX11-FAKE16-NEXT: v_perm_b32 v0, v1, v0, 0x5040100 ; encoding: [0x00,0x00,0x44,0xd6,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]99; GFX11-FAKE16-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v3, v5, v[6:8]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D ; encoding: [0x09,0x0f,0xe4,0xf0,0x00,0x00,0x00,0x08,0x02,0x03,0x05,0x06]100; GFX11-FAKE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]101; GFX11-FAKE16-NEXT: ; return to shader part epilog102;103; GFX12-TRUE16-LABEL: sample_d_3d:104; GFX12-TRUE16: ; %bb.0: ; %main_body105; GFX12-TRUE16-NEXT: v_mov_b16_e32 v3.h, v4.l ; encoding: [0x04,0x39,0x06,0x7f]106; GFX12-TRUE16-NEXT: v_mov_b16_e32 v0.h, v1.l ; encoding: [0x01,0x39,0x00,0x7f]107; GFX12-TRUE16-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v3, v[5:8]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D ; encoding: [0x02,0x40,0xce,0xe7,0x00,0x00,0x00,0x04,0x00,0x02,0x03,0x05]108; GFX12-TRUE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]109; GFX12-TRUE16-NEXT: ; return to shader part epilog110;111; GFX12-FAKE16-LABEL: sample_d_3d:112; GFX12-FAKE16: ; %bb.0: ; %main_body113; GFX12-FAKE16-NEXT: v_perm_b32 v3, v4, v3, 0x5040100 ; encoding: [0x03,0x00,0x44,0xd6,0x04,0x07,0xfe,0x03,0x00,0x01,0x04,0x05]114; GFX12-FAKE16-NEXT: v_perm_b32 v0, v1, v0, 0x5040100 ; encoding: [0x00,0x00,0x44,0xd6,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]115; GFX12-FAKE16-NEXT: image_sample_d_g16 v[0:3], [v0, v2, v3, v[5:8]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_3D ; encoding: [0x02,0x40,0xce,0xe7,0x00,0x00,0x00,0x04,0x00,0x02,0x03,0x05]116; GFX12-FAKE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]117; GFX12-FAKE16-NEXT: ; return to shader part epilog118main_body:119 %v = call <4 x float> @llvm.amdgcn.image.sample.d.3d.v4f32.f16.f32(i32 15, half %dsdh, half %dtdh, half %drdh, half %dsdv, half %dtdv, half %drdv, float %s, float %t, float %r, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)120 ret <4 x float> %v121}122 123define amdgpu_ps <4 x float> @sample_c_d_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, float %zcompare, half %dsdh, half %dsdv, float %s) {124; GFX10-LABEL: sample_c_d_1d:125; GFX10: ; %bb.0: ; %main_body126; GFX10-NEXT: image_sample_c_d_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x00]127; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]128; GFX10-NEXT: ; return to shader part epilog129;130; GFX11-LABEL: sample_c_d_1d:131; GFX11: ; %bb.0: ; %main_body132; GFX11-NEXT: image_sample_c_d_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x0f,0xe8,0xf0,0x00,0x00,0x00,0x08]133; GFX11-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]134; GFX11-NEXT: ; return to shader part epilog135;136; GFX12-LABEL: sample_c_d_1d:137; GFX12: ; %bb.0: ; %main_body138; GFX12-NEXT: image_sample_c_d_g16 v[0:3], [v0, v1, v2, v3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x80,0xce,0xe7,0x00,0x00,0x00,0x04,0x00,0x01,0x02,0x03]139; GFX12-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]140; GFX12-NEXT: ; return to shader part epilog141main_body:142 %v = call <4 x float> @llvm.amdgcn.image.sample.c.d.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)143 ret <4 x float> %v144}145 146define amdgpu_ps <4 x float> @sample_c_d_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) {147; GFX10-LABEL: sample_c_d_2d:148; GFX10: ; %bb.0: ; %main_body149; GFX10-NEXT: v_perm_b32 v3, v4, v3, 0x5040100 ; encoding: [0x03,0x00,0x44,0xd7,0x04,0x07,0xfe,0x03,0x00,0x01,0x04,0x05]150; GFX10-NEXT: v_perm_b32 v1, v2, v1, 0x5040100 ; encoding: [0x01,0x00,0x44,0xd7,0x02,0x03,0xfe,0x03,0x00,0x01,0x04,0x05]151; GFX10-NEXT: image_sample_c_d_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0xa8,0xf0,0x00,0x00,0x40,0x00,0x01,0x03,0x05,0x06]152; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]153; GFX10-NEXT: ; return to shader part epilog154;155; GFX11-TRUE16-LABEL: sample_c_d_2d:156; GFX11-TRUE16: ; %bb.0: ; %main_body157; GFX11-TRUE16-NEXT: v_mov_b16_e32 v3.h, v4.l ; encoding: [0x04,0x39,0x06,0x7f]158; GFX11-TRUE16-NEXT: v_mov_b16_e32 v1.h, v2.l ; encoding: [0x02,0x39,0x02,0x7f]159; GFX11-TRUE16-NEXT: image_sample_c_d_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x05,0x0f,0xe8,0xf0,0x00,0x00,0x00,0x08,0x01,0x03,0x05,0x06]160; GFX11-TRUE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]161; GFX11-TRUE16-NEXT: ; return to shader part epilog162;163; GFX11-FAKE16-LABEL: sample_c_d_2d:164; GFX11-FAKE16: ; %bb.0: ; %main_body165; GFX11-FAKE16-NEXT: v_perm_b32 v3, v4, v3, 0x5040100 ; encoding: [0x03,0x00,0x44,0xd6,0x04,0x07,0xfe,0x03,0x00,0x01,0x04,0x05]166; GFX11-FAKE16-NEXT: v_perm_b32 v1, v2, v1, 0x5040100 ; encoding: [0x01,0x00,0x44,0xd6,0x02,0x03,0xfe,0x03,0x00,0x01,0x04,0x05]167; GFX11-FAKE16-NEXT: image_sample_c_d_g16 v[0:3], [v0, v1, v3, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x05,0x0f,0xe8,0xf0,0x00,0x00,0x00,0x08,0x01,0x03,0x05,0x06]168; GFX11-FAKE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]169; GFX11-FAKE16-NEXT: ; return to shader part epilog170;171; GFX12-TRUE16-LABEL: sample_c_d_2d:172; GFX12-TRUE16: ; %bb.0: ; %main_body173; GFX12-TRUE16-NEXT: v_mov_b16_e32 v3.h, v4.l ; encoding: [0x04,0x39,0x06,0x7f]174; GFX12-TRUE16-NEXT: v_mov_b16_e32 v1.h, v2.l ; encoding: [0x02,0x39,0x02,0x7f]175; GFX12-TRUE16-NEXT: image_sample_c_d_g16 v[0:3], [v0, v1, v3, v[5:6]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x01,0x80,0xce,0xe7,0x00,0x00,0x00,0x04,0x00,0x01,0x03,0x05]176; GFX12-TRUE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]177; GFX12-TRUE16-NEXT: ; return to shader part epilog178;179; GFX12-FAKE16-LABEL: sample_c_d_2d:180; GFX12-FAKE16: ; %bb.0: ; %main_body181; GFX12-FAKE16-NEXT: v_perm_b32 v3, v4, v3, 0x5040100 ; encoding: [0x03,0x00,0x44,0xd6,0x04,0x07,0xfe,0x03,0x00,0x01,0x04,0x05]182; GFX12-FAKE16-NEXT: v_perm_b32 v1, v2, v1, 0x5040100 ; encoding: [0x01,0x00,0x44,0xd6,0x02,0x03,0xfe,0x03,0x00,0x01,0x04,0x05]183; GFX12-FAKE16-NEXT: image_sample_c_d_g16 v[0:3], [v0, v1, v3, v[5:6]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x01,0x80,0xce,0xe7,0x00,0x00,0x00,0x04,0x00,0x01,0x03,0x05]184; GFX12-FAKE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]185; GFX12-FAKE16-NEXT: ; return to shader part epilog186main_body:187 %v = call <4 x float> @llvm.amdgcn.image.sample.c.d.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)188 ret <4 x float> %v189}190 191define amdgpu_ps <4 x float> @sample_d_cl_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, half %dsdh, half %dsdv, float %s, float %clamp) {192; GFX10-LABEL: sample_d_cl_1d:193; GFX10: ; %bb.0: ; %main_body194; GFX10-NEXT: image_sample_d_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x00]195; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]196; GFX10-NEXT: ; return to shader part epilog197;198; GFX11-LABEL: sample_d_cl_1d:199; GFX11: ; %bb.0: ; %main_body200; GFX11-NEXT: image_sample_d_cl_g16 v[0:3], v[0:3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x0f,0x7c,0xf1,0x00,0x00,0x00,0x08]201; GFX11-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]202; GFX11-NEXT: ; return to shader part epilog203;204; GFX12-LABEL: sample_d_cl_1d:205; GFX12: ; %bb.0: ; %main_body206; GFX12-NEXT: image_sample_d_cl_g16 v[0:3], [v0, v1, v2, v3], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0xc0,0xd7,0xe7,0x00,0x00,0x00,0x04,0x00,0x01,0x02,0x03]207; GFX12-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]208; GFX12-NEXT: ; return to shader part epilog209main_body:210 %v = call <4 x float> @llvm.amdgcn.image.sample.d.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)211 ret <4 x float> %v212}213 214define amdgpu_ps <4 x float> @sample_d_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) {215; GFX10-LABEL: sample_d_cl_2d:216; GFX10: ; %bb.0: ; %main_body217; GFX10-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd7,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]218; GFX10-NEXT: v_perm_b32 v0, v1, v0, 0x5040100 ; encoding: [0x00,0x00,0x44,0xd7,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]219; GFX10-NEXT: image_sample_d_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x0b,0x0f,0x8c,0xf0,0x00,0x00,0x40,0x00,0x02,0x04,0x05,0x06]220; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]221; GFX10-NEXT: ; return to shader part epilog222;223; GFX11-TRUE16-LABEL: sample_d_cl_2d:224; GFX11-TRUE16: ; %bb.0: ; %main_body225; GFX11-TRUE16-NEXT: v_mov_b16_e32 v2.h, v3.l ; encoding: [0x03,0x39,0x04,0x7f]226; GFX11-TRUE16-NEXT: v_mov_b16_e32 v0.h, v1.l ; encoding: [0x01,0x39,0x00,0x7f]227; GFX11-TRUE16-NEXT: image_sample_d_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x05,0x0f,0x7c,0xf1,0x00,0x00,0x00,0x08,0x02,0x04,0x05,0x06]228; GFX11-TRUE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]229; GFX11-TRUE16-NEXT: ; return to shader part epilog230;231; GFX11-FAKE16-LABEL: sample_d_cl_2d:232; GFX11-FAKE16: ; %bb.0: ; %main_body233; GFX11-FAKE16-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd6,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]234; GFX11-FAKE16-NEXT: v_perm_b32 v0, v1, v0, 0x5040100 ; encoding: [0x00,0x00,0x44,0xd6,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]235; GFX11-FAKE16-NEXT: image_sample_d_cl_g16 v[0:3], [v0, v2, v4, v5, v6], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x05,0x0f,0x7c,0xf1,0x00,0x00,0x00,0x08,0x02,0x04,0x05,0x06]236; GFX11-FAKE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]237; GFX11-FAKE16-NEXT: ; return to shader part epilog238;239; GFX12-TRUE16-LABEL: sample_d_cl_2d:240; GFX12-TRUE16: ; %bb.0: ; %main_body241; GFX12-TRUE16-NEXT: v_mov_b16_e32 v2.h, v3.l ; encoding: [0x03,0x39,0x04,0x7f]242; GFX12-TRUE16-NEXT: v_mov_b16_e32 v0.h, v1.l ; encoding: [0x01,0x39,0x00,0x7f]243; GFX12-TRUE16-NEXT: image_sample_d_cl_g16 v[0:3], [v0, v2, v4, v[5:6]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x01,0xc0,0xd7,0xe7,0x00,0x00,0x00,0x04,0x00,0x02,0x04,0x05]244; GFX12-TRUE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]245; GFX12-TRUE16-NEXT: ; return to shader part epilog246;247; GFX12-FAKE16-LABEL: sample_d_cl_2d:248; GFX12-FAKE16: ; %bb.0: ; %main_body249; GFX12-FAKE16-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd6,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]250; GFX12-FAKE16-NEXT: v_perm_b32 v0, v1, v0, 0x5040100 ; encoding: [0x00,0x00,0x44,0xd6,0x01,0x01,0xfe,0x03,0x00,0x01,0x04,0x05]251; GFX12-FAKE16-NEXT: image_sample_d_cl_g16 v[0:3], [v0, v2, v4, v[5:6]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x01,0xc0,0xd7,0xe7,0x00,0x00,0x00,0x04,0x00,0x02,0x04,0x05]252; GFX12-FAKE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]253; GFX12-FAKE16-NEXT: ; return to shader part epilog254main_body:255 %v = call <4 x float> @llvm.amdgcn.image.sample.d.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)256 ret <4 x float> %v257}258 259define amdgpu_ps <4 x float> @sample_c_d_cl_1d(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, float %zcompare, half %dsdh, half %dsdv, float %s, float %clamp) {260; GFX10-LABEL: sample_c_d_cl_1d:261; GFX10: ; %bb.0: ; %main_body262; GFX10-NEXT: image_sample_c_d_cl_g16 v[0:3], v[0:4], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x01,0x0f,0xac,0xf0,0x00,0x00,0x40,0x00]263; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]264; GFX10-NEXT: ; return to shader part epilog265;266; GFX11-LABEL: sample_c_d_cl_1d:267; GFX11: ; %bb.0: ; %main_body268; GFX11-NEXT: image_sample_c_d_cl_g16 v[0:3], v[0:4], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x0f,0x50,0xf1,0x00,0x00,0x00,0x08]269; GFX11-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]270; GFX11-NEXT: ; return to shader part epilog271;272; GFX12-LABEL: sample_c_d_cl_1d:273; GFX12: ; %bb.0: ; %main_body274; GFX12-NEXT: image_sample_c_d_cl_g16 v[0:3], [v0, v1, v2, v[3:4]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x00,0xd5,0xe7,0x00,0x00,0x00,0x04,0x00,0x01,0x02,0x03]275; GFX12-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]276; GFX12-NEXT: ; return to shader part epilog277main_body:278 %v = call <4 x float> @llvm.amdgcn.image.sample.c.d.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)279 ret <4 x float> %v280}281 282define amdgpu_ps <4 x float> @sample_c_d_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) {283; GFX10-LABEL: sample_c_d_cl_2d:284; GFX10: ; %bb.0: ; %main_body285; GFX10-NEXT: v_mov_b32_e32 v8, v2 ; encoding: [0x02,0x03,0x10,0x7e]286; GFX10-NEXT: v_mov_b32_e32 v2, v0 ; encoding: [0x00,0x03,0x04,0x7e]287; GFX10-NEXT: v_perm_b32 v4, v4, v3, 0x5040100 ; encoding: [0x04,0x00,0x44,0xd7,0x04,0x07,0xfe,0x03,0x00,0x01,0x04,0x05]288; GFX10-NEXT: v_perm_b32 v3, v8, v1, 0x5040100 ; encoding: [0x03,0x00,0x44,0xd7,0x08,0x03,0xfe,0x03,0x00,0x01,0x04,0x05]289; GFX10-NEXT: image_sample_c_d_cl_g16 v[0:3], v[2:7], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x09,0x0f,0xac,0xf0,0x02,0x00,0x40,0x00]290; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]291; GFX10-NEXT: ; return to shader part epilog292;293; GFX11-TRUE16-LABEL: sample_c_d_cl_2d:294; GFX11-TRUE16: ; %bb.0: ; %main_body295; GFX11-TRUE16-NEXT: v_mov_b16_e32 v3.h, v4.l ; encoding: [0x04,0x39,0x06,0x7f]296; GFX11-TRUE16-NEXT: v_mov_b16_e32 v1.h, v2.l ; encoding: [0x02,0x39,0x02,0x7f]297; GFX11-TRUE16-NEXT: image_sample_c_d_cl_g16 v[0:3], [v0, v1, v3, v5, v[6:7]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x05,0x0f,0x50,0xf1,0x00,0x00,0x00,0x08,0x01,0x03,0x05,0x06]298; GFX11-TRUE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]299; GFX11-TRUE16-NEXT: ; return to shader part epilog300;301; GFX11-FAKE16-LABEL: sample_c_d_cl_2d:302; GFX11-FAKE16: ; %bb.0: ; %main_body303; GFX11-FAKE16-NEXT: v_perm_b32 v3, v4, v3, 0x5040100 ; encoding: [0x03,0x00,0x44,0xd6,0x04,0x07,0xfe,0x03,0x00,0x01,0x04,0x05]304; GFX11-FAKE16-NEXT: v_perm_b32 v1, v2, v1, 0x5040100 ; encoding: [0x01,0x00,0x44,0xd6,0x02,0x03,0xfe,0x03,0x00,0x01,0x04,0x05]305; GFX11-FAKE16-NEXT: image_sample_c_d_cl_g16 v[0:3], [v0, v1, v3, v5, v[6:7]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x05,0x0f,0x50,0xf1,0x00,0x00,0x00,0x08,0x01,0x03,0x05,0x06]306; GFX11-FAKE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]307; GFX11-FAKE16-NEXT: ; return to shader part epilog308;309; GFX12-TRUE16-LABEL: sample_c_d_cl_2d:310; GFX12-TRUE16: ; %bb.0: ; %main_body311; GFX12-TRUE16-NEXT: v_mov_b16_e32 v3.h, v4.l ; encoding: [0x04,0x39,0x06,0x7f]312; GFX12-TRUE16-NEXT: v_mov_b16_e32 v1.h, v2.l ; encoding: [0x02,0x39,0x02,0x7f]313; GFX12-TRUE16-NEXT: image_sample_c_d_cl_g16 v[0:3], [v0, v1, v3, v[5:7]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x01,0x00,0xd5,0xe7,0x00,0x00,0x00,0x04,0x00,0x01,0x03,0x05]314; GFX12-TRUE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]315; GFX12-TRUE16-NEXT: ; return to shader part epilog316;317; GFX12-FAKE16-LABEL: sample_c_d_cl_2d:318; GFX12-FAKE16: ; %bb.0: ; %main_body319; GFX12-FAKE16-NEXT: v_perm_b32 v3, v4, v3, 0x5040100 ; encoding: [0x03,0x00,0x44,0xd6,0x04,0x07,0xfe,0x03,0x00,0x01,0x04,0x05]320; GFX12-FAKE16-NEXT: v_perm_b32 v1, v2, v1, 0x5040100 ; encoding: [0x01,0x00,0x44,0xd6,0x02,0x03,0xfe,0x03,0x00,0x01,0x04,0x05]321; GFX12-FAKE16-NEXT: image_sample_c_d_cl_g16 v[0:3], [v0, v1, v3, v[5:7]], s[0:7], s[8:11] dmask:0xf dim:SQ_RSRC_IMG_2D ; encoding: [0x01,0x00,0xd5,0xe7,0x00,0x00,0x00,0x04,0x00,0x01,0x03,0x05]322; GFX12-FAKE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]323; GFX12-FAKE16-NEXT: ; return to shader part epilog324main_body:325 %v = call <4 x float> @llvm.amdgcn.image.sample.c.d.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)326 ret <4 x float> %v327}328 329define amdgpu_ps float @sample_c_d_o_2darray_V1(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, i32 %offset, float %zcompare, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, float %slice) {330; GFX10-LABEL: sample_c_d_o_2darray_V1:331; GFX10: ; %bb.0: ; %main_body332; GFX10-NEXT: v_mov_b32_e32 v9, v3 ; encoding: [0x03,0x03,0x12,0x7e]333; GFX10-NEXT: v_mov_b32_e32 v10, v2 ; encoding: [0x02,0x03,0x14,0x7e]334; GFX10-NEXT: v_mov_b32_e32 v3, v1 ; encoding: [0x01,0x03,0x06,0x7e]335; GFX10-NEXT: v_mov_b32_e32 v2, v0 ; encoding: [0x00,0x03,0x04,0x7e]336; GFX10-NEXT: v_perm_b32 v5, v5, v4, 0x5040100 ; encoding: [0x05,0x00,0x44,0xd7,0x05,0x09,0xfe,0x03,0x00,0x01,0x04,0x05]337; GFX10-NEXT: v_perm_b32 v4, v9, v10, 0x5040100 ; encoding: [0x04,0x00,0x44,0xd7,0x09,0x15,0xfe,0x03,0x00,0x01,0x04,0x05]338; GFX10-NEXT: image_sample_c_d_o_g16 v0, v[2:8], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x29,0x04,0xe8,0xf0,0x02,0x00,0x40,0x00]339; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]340; GFX10-NEXT: ; return to shader part epilog341;342; GFX11-TRUE16-LABEL: sample_c_d_o_2darray_V1:343; GFX11-TRUE16: ; %bb.0: ; %main_body344; GFX11-TRUE16-NEXT: v_mov_b16_e32 v4.h, v5.l ; encoding: [0x05,0x39,0x08,0x7f]345; GFX11-TRUE16-NEXT: v_mov_b16_e32 v2.h, v3.l ; encoding: [0x03,0x39,0x04,0x7f]346; GFX11-TRUE16-NEXT: image_sample_c_d_o_g16 v0, [v0, v1, v2, v4, v[6:8]], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x15,0x04,0xf0,0xf0,0x00,0x00,0x00,0x08,0x01,0x02,0x04,0x06]347; GFX11-TRUE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]348; GFX11-TRUE16-NEXT: ; return to shader part epilog349;350; GFX11-FAKE16-LABEL: sample_c_d_o_2darray_V1:351; GFX11-FAKE16: ; %bb.0: ; %main_body352; GFX11-FAKE16-NEXT: v_perm_b32 v4, v5, v4, 0x5040100 ; encoding: [0x04,0x00,0x44,0xd6,0x05,0x09,0xfe,0x03,0x00,0x01,0x04,0x05]353; GFX11-FAKE16-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd6,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]354; GFX11-FAKE16-NEXT: image_sample_c_d_o_g16 v0, [v0, v1, v2, v4, v[6:8]], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x15,0x04,0xf0,0xf0,0x00,0x00,0x00,0x08,0x01,0x02,0x04,0x06]355; GFX11-FAKE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]356; GFX11-FAKE16-NEXT: ; return to shader part epilog357;358; GFX12-TRUE16-LABEL: sample_c_d_o_2darray_V1:359; GFX12-TRUE16: ; %bb.0: ; %main_body360; GFX12-TRUE16-NEXT: v_mov_b32_e32 v9, v5 ; encoding: [0x05,0x03,0x12,0x7e]361; GFX12-TRUE16-NEXT: v_mov_b16_e32 v5.l, v4.l ; encoding: [0x04,0x39,0x0a,0x7e]362; GFX12-TRUE16-NEXT: v_mov_b16_e32 v2.h, v3.l ; encoding: [0x03,0x39,0x04,0x7f]363; GFX12-TRUE16-NEXT: s_delay_alu instid0(VALU_DEP_3) ; encoding: [0x03,0x00,0x87,0xbf]364; GFX12-TRUE16-NEXT: v_mov_b16_e32 v5.h, v9.l ; encoding: [0x09,0x39,0x0a,0x7f]365; GFX12-TRUE16-NEXT: image_sample_c_d_o_g16 v0, [v0, v1, v2, v[5:8]], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x05,0x00,0x0f,0xe5,0x00,0x00,0x00,0x04,0x00,0x01,0x02,0x05]366; GFX12-TRUE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]367; GFX12-TRUE16-NEXT: ; return to shader part epilog368;369; GFX12-FAKE16-LABEL: sample_c_d_o_2darray_V1:370; GFX12-FAKE16: ; %bb.0: ; %main_body371; GFX12-FAKE16-NEXT: v_perm_b32 v5, v5, v4, 0x5040100 ; encoding: [0x05,0x00,0x44,0xd6,0x05,0x09,0xfe,0x03,0x00,0x01,0x04,0x05]372; GFX12-FAKE16-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd6,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]373; GFX12-FAKE16-NEXT: image_sample_c_d_o_g16 v0, [v0, v1, v2, v[5:8]], s[0:7], s[8:11] dmask:0x4 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x05,0x00,0x0f,0xe5,0x00,0x00,0x00,0x04,0x00,0x01,0x02,0x05]374; GFX12-FAKE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]375; GFX12-FAKE16-NEXT: ; return to shader part epilog376main_body:377 %v = call float @llvm.amdgcn.image.sample.c.d.o.2darray.f16.f32.f32(i32 4, i32 %offset, float %zcompare, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, float %slice, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)378 ret float %v379}380 381define amdgpu_ps <2 x float> @sample_c_d_o_2darray_V2(<8 x i32> inreg %rsrc, <4 x i32> inreg %samp, i32 %offset, float %zcompare, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, float %slice) {382; GFX10-LABEL: sample_c_d_o_2darray_V2:383; GFX10: ; %bb.0: ; %main_body384; GFX10-NEXT: v_mov_b32_e32 v9, v3 ; encoding: [0x03,0x03,0x12,0x7e]385; GFX10-NEXT: v_mov_b32_e32 v10, v2 ; encoding: [0x02,0x03,0x14,0x7e]386; GFX10-NEXT: v_mov_b32_e32 v3, v1 ; encoding: [0x01,0x03,0x06,0x7e]387; GFX10-NEXT: v_mov_b32_e32 v2, v0 ; encoding: [0x00,0x03,0x04,0x7e]388; GFX10-NEXT: v_perm_b32 v5, v5, v4, 0x5040100 ; encoding: [0x05,0x00,0x44,0xd7,0x05,0x09,0xfe,0x03,0x00,0x01,0x04,0x05]389; GFX10-NEXT: v_perm_b32 v4, v9, v10, 0x5040100 ; encoding: [0x04,0x00,0x44,0xd7,0x09,0x15,0xfe,0x03,0x00,0x01,0x04,0x05]390; GFX10-NEXT: image_sample_c_d_o_g16 v[0:1], v[2:8], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x29,0x06,0xe8,0xf0,0x02,0x00,0x40,0x00]391; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]392; GFX10-NEXT: ; return to shader part epilog393;394; GFX11-TRUE16-LABEL: sample_c_d_o_2darray_V2:395; GFX11-TRUE16: ; %bb.0: ; %main_body396; GFX11-TRUE16-NEXT: v_mov_b16_e32 v4.h, v5.l ; encoding: [0x05,0x39,0x08,0x7f]397; GFX11-TRUE16-NEXT: v_mov_b16_e32 v2.h, v3.l ; encoding: [0x03,0x39,0x04,0x7f]398; GFX11-TRUE16-NEXT: image_sample_c_d_o_g16 v[0:1], [v0, v1, v2, v4, v[6:8]], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x15,0x06,0xf0,0xf0,0x00,0x00,0x00,0x08,0x01,0x02,0x04,0x06]399; GFX11-TRUE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]400; GFX11-TRUE16-NEXT: ; return to shader part epilog401;402; GFX11-FAKE16-LABEL: sample_c_d_o_2darray_V2:403; GFX11-FAKE16: ; %bb.0: ; %main_body404; GFX11-FAKE16-NEXT: v_perm_b32 v4, v5, v4, 0x5040100 ; encoding: [0x04,0x00,0x44,0xd6,0x05,0x09,0xfe,0x03,0x00,0x01,0x04,0x05]405; GFX11-FAKE16-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd6,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]406; GFX11-FAKE16-NEXT: image_sample_c_d_o_g16 v[0:1], [v0, v1, v2, v4, v[6:8]], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x15,0x06,0xf0,0xf0,0x00,0x00,0x00,0x08,0x01,0x02,0x04,0x06]407; GFX11-FAKE16-NEXT: s_waitcnt vmcnt(0) ; encoding: [0xf7,0x03,0x89,0xbf]408; GFX11-FAKE16-NEXT: ; return to shader part epilog409;410; GFX12-TRUE16-LABEL: sample_c_d_o_2darray_V2:411; GFX12-TRUE16: ; %bb.0: ; %main_body412; GFX12-TRUE16-NEXT: v_mov_b32_e32 v9, v5 ; encoding: [0x05,0x03,0x12,0x7e]413; GFX12-TRUE16-NEXT: v_mov_b16_e32 v5.l, v4.l ; encoding: [0x04,0x39,0x0a,0x7e]414; GFX12-TRUE16-NEXT: v_mov_b16_e32 v2.h, v3.l ; encoding: [0x03,0x39,0x04,0x7f]415; GFX12-TRUE16-NEXT: s_delay_alu instid0(VALU_DEP_3) ; encoding: [0x03,0x00,0x87,0xbf]416; GFX12-TRUE16-NEXT: v_mov_b16_e32 v5.h, v9.l ; encoding: [0x09,0x39,0x0a,0x7f]417; GFX12-TRUE16-NEXT: image_sample_c_d_o_g16 v[0:1], [v0, v1, v2, v[5:8]], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x05,0x00,0x8f,0xe5,0x00,0x00,0x00,0x04,0x00,0x01,0x02,0x05]418; GFX12-TRUE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]419; GFX12-TRUE16-NEXT: ; return to shader part epilog420;421; GFX12-FAKE16-LABEL: sample_c_d_o_2darray_V2:422; GFX12-FAKE16: ; %bb.0: ; %main_body423; GFX12-FAKE16-NEXT: v_perm_b32 v5, v5, v4, 0x5040100 ; encoding: [0x05,0x00,0x44,0xd6,0x05,0x09,0xfe,0x03,0x00,0x01,0x04,0x05]424; GFX12-FAKE16-NEXT: v_perm_b32 v2, v3, v2, 0x5040100 ; encoding: [0x02,0x00,0x44,0xd6,0x03,0x05,0xfe,0x03,0x00,0x01,0x04,0x05]425; GFX12-FAKE16-NEXT: image_sample_c_d_o_g16 v[0:1], [v0, v1, v2, v[5:8]], s[0:7], s[8:11] dmask:0x6 dim:SQ_RSRC_IMG_2D_ARRAY ; encoding: [0x05,0x00,0x8f,0xe5,0x00,0x00,0x00,0x04,0x00,0x01,0x02,0x05]426; GFX12-FAKE16-NEXT: s_wait_samplecnt 0x0 ; encoding: [0x00,0x00,0xc2,0xbf]427; GFX12-FAKE16-NEXT: ; return to shader part epilog428main_body:429 %v = call <2 x float> @llvm.amdgcn.image.sample.c.d.o.2darray.v2f32.f16.f32(i32 6, i32 %offset, float %zcompare, half %dsdh, half %dtdh, half %dsdv, half %dtdv, float %s, float %t, float %slice, <8 x i32> %rsrc, <4 x i32> %samp, i1 0, i32 0, i32 0)430 ret <2 x float> %v431}432 433declare <4 x float> @llvm.amdgcn.image.sample.d.1d.v4f32.f16.f32(i32, half, half, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1434declare <4 x float> @llvm.amdgcn.image.sample.d.2d.v4f32.f16.f32(i32, half, half, half, half, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1435declare <4 x float> @llvm.amdgcn.image.sample.d.3d.v4f32.f16.f32(i32, half, half, half, half, half, half, float, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1436declare <4 x float> @llvm.amdgcn.image.sample.c.d.1d.v4f32.f16.f32(i32, float, half, half, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1437declare <4 x float> @llvm.amdgcn.image.sample.c.d.2d.v4f32.f16.f32(i32, float, half, half, half, half, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1438declare <4 x float> @llvm.amdgcn.image.sample.d.cl.1d.v4f32.f16.f32(i32, half, half, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1439declare <4 x float> @llvm.amdgcn.image.sample.d.cl.2d.v4f32.f16.f32(i32, half, half, half, half, float, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1440declare <4 x float> @llvm.amdgcn.image.sample.c.d.cl.1d.v4f32.f16.f32(i32, float, half, half, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1441declare <4 x float> @llvm.amdgcn.image.sample.c.d.cl.2d.v4f32.f16.f32(i32, float, half, half, half, half, float, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1442 443declare float @llvm.amdgcn.image.sample.c.d.o.2darray.f16.f32.f32(i32, i32, float, half, half, half, half, float, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1444declare <2 x float> @llvm.amdgcn.image.sample.c.d.o.2darray.v2f32.f16.f32(i32, i32, float, half, half, half, half, float, float, float, <8 x i32>, <4 x i32>, i1, i32, i32) #1445 446attributes #0 = { nounwind }447attributes #1 = { nounwind readonly }448attributes #2 = { nounwind readnone }449