mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-22 18:54:02 +01:00
c4a674e21e
the compilation time and there is no case for which we see any improvement in performance. This patch removes this pass and its associated test cases from the tree. Differential Revision: https://reviews.llvm.org/D101313 Change-Id: I0599169a7609c19a887f8d847a71e664030cc141
3553 lines
155 KiB
LLVM
3553 lines
155 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -march=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefixes=VERDE %s
|
|
; RUN: llc -march=amdgcn -mcpu=fiji -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefixes=FIJI %s
|
|
; RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefixes=GFX6789 %s
|
|
; RUN: llc -march=amdgcn -mcpu=gfx900 -mattr=-enable-prt-strict-null -verify-machineinstrs < %s | FileCheck -check-prefixes=NOPRT %s
|
|
; RUN: llc -march=amdgcn -mcpu=gfx1010 -verify-machineinstrs -show-mc-encoding < %s | FileCheck -check-prefixes=GFX10 %s
|
|
|
|
define amdgpu_ps <4 x float> @load_1d(<8 x i32> inreg %rsrc, i32 %s) {
|
|
; VERDE-LABEL: load_1d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i32(i32 15, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1d_tfe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s) {
|
|
; VERDE-LABEL: load_1d_tfe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v5, s[0:7] dmask:0xf unorm tfe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_tfe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v5, s[0:7] dmask:0xf unorm tfe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_tfe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v10
|
|
; GFX6789-NEXT: image_load v[0:4], v5, s[0:7] dmask:0xf unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v6, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_tfe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v0, s[0:7] dmask:0xf unorm tfe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_tfe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, 0 ; encoding: [0x80,0x02,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v6 ; encoding: [0x06,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, v6 ; encoding: [0x06,0x03,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v6 ; encoding: [0x06,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v6 ; encoding: [0x06,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v6 ; encoding: [0x06,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v7 ; encoding: [0x07,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v8 ; encoding: [0x08,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v9 ; encoding: [0x09,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v10 ; encoding: [0x0a,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v5, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm tfe ; encoding: [0x00,0x1f,0x01,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v6, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x06,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.1d.v4f32i32.i32(i32 15, i32 %s, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1d_lwe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s) {
|
|
; VERDE-LABEL: load_1d_lwe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v5, s[0:7] dmask:0xf unorm lwe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_lwe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v5, s[0:7] dmask:0xf unorm lwe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_lwe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v10
|
|
; GFX6789-NEXT: image_load v[0:4], v5, s[0:7] dmask:0xf unorm lwe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v6, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_lwe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v0, s[0:7] dmask:0xf unorm lwe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_lwe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, 0 ; encoding: [0x80,0x02,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v6 ; encoding: [0x06,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, v6 ; encoding: [0x06,0x03,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v6 ; encoding: [0x06,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v6 ; encoding: [0x06,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v6 ; encoding: [0x06,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v7 ; encoding: [0x07,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v8 ; encoding: [0x08,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v9 ; encoding: [0x09,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v10 ; encoding: [0x0a,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v5, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm lwe ; encoding: [0x00,0x1f,0x02,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v6, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x06,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>, i32} @llvm.amdgcn.image.load.1d.v4f32i32.i32(i32 15, i32 %s, <8 x i32> %rsrc, i32 2, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_2d(<8 x i32> inreg %rsrc, i32 %s, i32 %t) {
|
|
; VERDE-LABEL: load_2d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_2d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_2d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_2d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_2d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm ; encoding: [0x08,0x1f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.2d.v4f32.i32(i32 15, i32 %s, i32 %t, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_2d_tfe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %t) {
|
|
; VERDE-LABEL: load_2d_tfe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v[5:6], s[0:7] dmask:0xf unorm tfe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_2d_tfe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v[5:6], s[0:7] dmask:0xf unorm tfe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_2d_tfe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v11
|
|
; GFX6789-NEXT: image_load v[0:4], v[5:6], s[0:7] dmask:0xf unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v7, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_2d_tfe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v[0:1], s[0:7] dmask:0xf unorm tfe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_2d_tfe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, 0 ; encoding: [0x80,0x02,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, v7 ; encoding: [0x07,0x03,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v7 ; encoding: [0x07,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v7 ; encoding: [0x07,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v7 ; encoding: [0x07,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v7 ; encoding: [0x07,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v8 ; encoding: [0x08,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v9 ; encoding: [0x09,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v10 ; encoding: [0x0a,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v11 ; encoding: [0x0b,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v[5:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm tfe ; encoding: [0x08,0x1f,0x01,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v7, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x07,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.2d.v4f32i32.i32(i32 15, i32 %s, i32 %t, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_3d(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %r) {
|
|
; VERDE-LABEL: load_3d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_3d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_3d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_3d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_3d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm ; encoding: [0x10,0x1f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.3d.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %r, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_3d_tfe_lwe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %t, i32 %r) {
|
|
; VERDE-LABEL: load_3d_tfe_lwe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v7, v2
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe lwe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_3d_tfe_lwe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v7, v2
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe lwe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_3d_tfe_lwe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v2
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v12, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v11
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v12
|
|
; GFX6789-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe lwe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v8, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_3d_tfe_lwe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v[0:2], s[0:7] dmask:0xf unorm tfe lwe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_3d_tfe_lwe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, 0 ; encoding: [0x80,0x02,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v2 ; encoding: [0x02,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v8 ; encoding: [0x08,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v8 ; encoding: [0x08,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v8 ; encoding: [0x08,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v12, v8 ; encoding: [0x08,0x03,0x18,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v8 ; encoding: [0x08,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v9 ; encoding: [0x09,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v10 ; encoding: [0x0a,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v11 ; encoding: [0x0b,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v12 ; encoding: [0x0c,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm tfe lwe ; encoding: [0x10,0x1f,0x03,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v8, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x08,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.3d.v4f32i32.i32(i32 15, i32 %s, i32 %t, i32 %r, <8 x i32> %rsrc, i32 3, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_cube(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %slice) {
|
|
; VERDE-LABEL: load_cube:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_cube:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_cube:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_cube:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_cube:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm ; encoding: [0x18,0x1f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.cube.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %slice, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_cube_lwe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %t, i32 %slice) {
|
|
; VERDE-LABEL: load_cube_lwe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v7, v2
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm lwe da
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_cube_lwe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v7, v2
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm lwe da
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_cube_lwe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v2
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v12, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v11
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v12
|
|
; GFX6789-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm lwe da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v8, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_cube_lwe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v[0:2], s[0:7] dmask:0xf unorm lwe da
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_cube_lwe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, 0 ; encoding: [0x80,0x02,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v2 ; encoding: [0x02,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v8 ; encoding: [0x08,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v8 ; encoding: [0x08,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v8 ; encoding: [0x08,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v12, v8 ; encoding: [0x08,0x03,0x18,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v8 ; encoding: [0x08,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v9 ; encoding: [0x09,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v10 ; encoding: [0x0a,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v11 ; encoding: [0x0b,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v12 ; encoding: [0x0c,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm lwe ; encoding: [0x18,0x1f,0x02,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v8, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x08,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.cube.v4f32i32.i32(i32 15, i32 %s, i32 %t, i32 %slice, <8 x i32> %rsrc, i32 2, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1darray(<8 x i32> inreg %rsrc, i32 %s, i32 %slice) {
|
|
; VERDE-LABEL: load_1darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm ; encoding: [0x20,0x1f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.1darray.v4f32.i32(i32 15, i32 %s, i32 %slice, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1darray_tfe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %slice) {
|
|
; VERDE-LABEL: load_1darray_tfe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v[5:6], s[0:7] dmask:0xf unorm tfe da
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1darray_tfe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v[5:6], s[0:7] dmask:0xf unorm tfe da
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1darray_tfe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v11
|
|
; GFX6789-NEXT: image_load v[0:4], v[5:6], s[0:7] dmask:0xf unorm tfe da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v7, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1darray_tfe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v[0:1], s[0:7] dmask:0xf unorm tfe da
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1darray_tfe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, 0 ; encoding: [0x80,0x02,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, v7 ; encoding: [0x07,0x03,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v7 ; encoding: [0x07,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v7 ; encoding: [0x07,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v7 ; encoding: [0x07,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v7 ; encoding: [0x07,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v8 ; encoding: [0x08,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v9 ; encoding: [0x09,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v10 ; encoding: [0x0a,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v11 ; encoding: [0x0b,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v[5:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm tfe ; encoding: [0x20,0x1f,0x01,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v7, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x07,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.1darray.v4f32i32.i32(i32 15, i32 %s, i32 %slice, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_2darray(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %slice) {
|
|
; VERDE-LABEL: load_2darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_2darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_2darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_2darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_2darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm ; encoding: [0x28,0x1f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.2darray.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %slice, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_2darray_lwe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %t, i32 %slice) {
|
|
; VERDE-LABEL: load_2darray_lwe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v7, v2
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm lwe da
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_2darray_lwe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v7, v2
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm lwe da
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_2darray_lwe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v2
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v12, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v11
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v12
|
|
; GFX6789-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm lwe da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v8, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_2darray_lwe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v[0:2], s[0:7] dmask:0xf unorm lwe da
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_2darray_lwe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, 0 ; encoding: [0x80,0x02,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v2 ; encoding: [0x02,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v8 ; encoding: [0x08,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v8 ; encoding: [0x08,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v8 ; encoding: [0x08,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v12, v8 ; encoding: [0x08,0x03,0x18,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v8 ; encoding: [0x08,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v9 ; encoding: [0x09,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v10 ; encoding: [0x0a,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v11 ; encoding: [0x0b,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v12 ; encoding: [0x0c,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm lwe ; encoding: [0x28,0x1f,0x02,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v8, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x08,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.2darray.v4f32i32.i32(i32 15, i32 %s, i32 %t, i32 %slice, <8 x i32> %rsrc, i32 2, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_2dmsaa(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %fragid) {
|
|
; VERDE-LABEL: load_2dmsaa:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_2dmsaa:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_2dmsaa:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_2dmsaa:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_2dmsaa:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v[0:2], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm ; encoding: [0x30,0x1f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.2dmsaa.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_2dmsaa_both(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %t, i32 %fragid) {
|
|
; VERDE-LABEL: load_2dmsaa_both:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v7, v2
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe lwe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_2dmsaa_both:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v7, v2
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe lwe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_2dmsaa_both:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v2
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v12, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v11
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v12
|
|
; GFX6789-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe lwe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v8, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_2dmsaa_both:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v[0:2], s[0:7] dmask:0xf unorm tfe lwe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_2dmsaa_both:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, 0 ; encoding: [0x80,0x02,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v2 ; encoding: [0x02,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v8 ; encoding: [0x08,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v8 ; encoding: [0x08,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v8 ; encoding: [0x08,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v12, v8 ; encoding: [0x08,0x03,0x18,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v8 ; encoding: [0x08,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v9 ; encoding: [0x09,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v10 ; encoding: [0x0a,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v11 ; encoding: [0x0b,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v12 ; encoding: [0x0c,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm tfe lwe ; encoding: [0x30,0x1f,0x03,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v8, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x08,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.2dmsaa.v4f32i32.i32(i32 15, i32 %s, i32 %t, i32 %fragid, <8 x i32> %rsrc, i32 3, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_2darraymsaa(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %slice, i32 %fragid) {
|
|
; VERDE-LABEL: load_2darraymsaa:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_2darraymsaa:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_2darraymsaa:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_2darraymsaa:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_2darraymsaa:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v[0:3], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm ; encoding: [0x38,0x1f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.2darraymsaa.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %slice, i32 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_2darraymsaa_tfe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %t, i32 %slice, i32 %fragid) {
|
|
; VERDE-LABEL: load_2darraymsaa_tfe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v8, v3
|
|
; VERDE-NEXT: v_mov_b32_e32 v7, v2
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load v[0:4], v[5:8], s[0:7] dmask:0xf unorm tfe da
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_2darraymsaa_tfe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v8, v3
|
|
; FIJI-NEXT: v_mov_b32_e32 v7, v2
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load v[0:4], v[5:8], s[0:7] dmask:0xf unorm tfe da
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_2darraymsaa_tfe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, v3
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v2
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v12, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v13, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v11
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v12
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v13
|
|
; GFX6789-NEXT: image_load v[0:4], v[5:8], s[0:7] dmask:0xf unorm tfe da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v9, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_2darraymsaa_tfe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load v[0:4], v[0:3], s[0:7] dmask:0xf unorm tfe da
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_2darraymsaa_tfe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, 0 ; encoding: [0x80,0x02,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, v3 ; encoding: [0x03,0x03,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v2 ; encoding: [0x02,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v9 ; encoding: [0x09,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v9 ; encoding: [0x09,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v12, v9 ; encoding: [0x09,0x03,0x18,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v13, v9 ; encoding: [0x09,0x03,0x1a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v9 ; encoding: [0x09,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v10 ; encoding: [0x0a,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v11 ; encoding: [0x0b,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v12 ; encoding: [0x0c,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v13 ; encoding: [0x0d,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load v[0:4], v[5:8], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm tfe ; encoding: [0x38,0x1f,0x01,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v9, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x09,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.2darraymsaa.v4f32i32.i32(i32 15, i32 %s, i32 %t, i32 %slice, i32 %fragid, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_mip_1d(<8 x i32> inreg %rsrc, i32 %s, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_1d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_1d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_1d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_1d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_1d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load_mip v[0:3], v[0:1], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1f,0x04,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.mip.1d.v4f32.i32(i32 15, i32 %s, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_mip_1d_lwe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_1d_lwe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load_mip v[0:4], v[5:6], s[0:7] dmask:0xf unorm lwe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_1d_lwe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load_mip v[0:4], v[5:6], s[0:7] dmask:0xf unorm lwe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_1d_lwe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v11
|
|
; GFX6789-NEXT: image_load_mip v[0:4], v[5:6], s[0:7] dmask:0xf unorm lwe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v7, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_1d_lwe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load_mip v[0:4], v[0:1], s[0:7] dmask:0xf unorm lwe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_1d_lwe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, 0 ; encoding: [0x80,0x02,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, v7 ; encoding: [0x07,0x03,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v7 ; encoding: [0x07,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v7 ; encoding: [0x07,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v7 ; encoding: [0x07,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v7 ; encoding: [0x07,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v8 ; encoding: [0x08,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v9 ; encoding: [0x09,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v10 ; encoding: [0x0a,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v11 ; encoding: [0x0b,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load_mip v[0:4], v[5:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm lwe ; encoding: [0x00,0x1f,0x06,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v7, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x07,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.mip.1d.v4f32i32.i32(i32 15, i32 %s, i32 %mip, <8 x i32> %rsrc, i32 2, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_mip_2d(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_2d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_2d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_2d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_2d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_2d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm ; encoding: [0x08,0x1f,0x04,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.mip.2d.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_mip_2d_tfe(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s, i32 %t, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_2d_tfe:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v5, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v7, v2
|
|
; VERDE-NEXT: v_mov_b32_e32 v6, v1
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: image_load_mip v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_2d_tfe:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v5, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v7, v2
|
|
; FIJI-NEXT: v_mov_b32_e32 v6, v1
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: image_load_mip v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v4, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_2d_tfe:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v2
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v1
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v9, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v10, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v11, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v12, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v8
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v9
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v10
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v11
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v12
|
|
; GFX6789-NEXT: image_load_mip v[0:4], v[5:7], s[0:7] dmask:0xf unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v8, v4, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_2d_tfe:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: image_load_mip v[0:4], v[0:2], s[0:7] dmask:0xf unorm tfe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v5, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v5, v4, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_2d_tfe:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, 0 ; encoding: [0x80,0x02,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v2 ; encoding: [0x02,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v1 ; encoding: [0x01,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v0 ; encoding: [0x00,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v9, v8 ; encoding: [0x08,0x03,0x12,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v10, v8 ; encoding: [0x08,0x03,0x14,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v11, v8 ; encoding: [0x08,0x03,0x16,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v12, v8 ; encoding: [0x08,0x03,0x18,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v8 ; encoding: [0x08,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v9 ; encoding: [0x09,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v10 ; encoding: [0x0a,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v11 ; encoding: [0x0b,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v12 ; encoding: [0x0c,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load_mip v[0:4], v[5:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm tfe ; encoding: [0x08,0x1f,0x05,0xf0,0x05,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v8, v4, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x08,0x04,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.mip.2d.v4f32i32.i32(i32 15, i32 %s, i32 %t, i32 %mip, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps float @load_1d_V2_tfe_dmask0(<8 x i32> inreg %rsrc, i32 %s) {
|
|
; VERDE-LABEL: load_1d_V2_tfe_dmask0:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v1
|
|
; VERDE-NEXT: image_load v[1:2], v0, s[0:7] dmask:0x1 unorm tfe
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, v2
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_V2_tfe_dmask0:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v1
|
|
; FIJI-NEXT: image_load v[1:2], v0, s[0:7] dmask:0x1 unorm tfe
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, v2
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_V2_tfe_dmask0:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v1
|
|
; GFX6789-NEXT: image_load v[1:2], v0, s[0:7] dmask:0x1 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v2
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_V2_tfe_dmask0:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v1, 0
|
|
; NOPRT-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x1 unorm tfe
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: v_mov_b32_e32 v0, v1
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_V2_tfe_dmask0:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, 0 ; encoding: [0x80,0x02,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v1 ; encoding: [0x01,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: image_load v[1:2], v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm tfe ; encoding: [0x00,0x11,0x01,0xf0,0x00,0x01,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v2 ; encoding: [0x02,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<2 x float>,i32} @llvm.amdgcn.image.load.1d.v2f32i32.i32(i32 0, i32 %s, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.err = extractvalue {<2 x float>, i32} %v, 1
|
|
%vv = bitcast i32 %v.err to float
|
|
ret float %vv
|
|
}
|
|
|
|
define amdgpu_ps float @load_1d_V1_tfe_dmask0(<8 x i32> inreg %rsrc, i32 %s) {
|
|
; VERDE-LABEL: load_1d_V1_tfe_dmask0:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v1
|
|
; VERDE-NEXT: image_load v[1:2], v0, s[0:7] dmask:0x1 unorm tfe
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, v2
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_V1_tfe_dmask0:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v1
|
|
; FIJI-NEXT: image_load v[1:2], v0, s[0:7] dmask:0x1 unorm tfe
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, v2
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_V1_tfe_dmask0:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v1
|
|
; GFX6789-NEXT: image_load v[1:2], v0, s[0:7] dmask:0x1 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v2
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_V1_tfe_dmask0:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v1, 0
|
|
; NOPRT-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x1 unorm tfe
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: v_mov_b32_e32 v0, v1
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_V1_tfe_dmask0:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, 0 ; encoding: [0x80,0x02,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v1 ; encoding: [0x01,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: image_load v[1:2], v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm tfe ; encoding: [0x00,0x11,0x01,0xf0,0x00,0x01,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v2 ; encoding: [0x02,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {float,i32} @llvm.amdgcn.image.load.1d.f32i32.i32(i32 0, i32 %s, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.err = extractvalue {float, i32} %v, 1
|
|
%vv = bitcast i32 %v.err to float
|
|
ret float %vv
|
|
}
|
|
|
|
define amdgpu_ps float @load_mip_2d_tfe_dmask0(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_2d_tfe_dmask0:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v3
|
|
; VERDE-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, v4
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_2d_tfe_dmask0:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v3
|
|
; FIJI-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, v4
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_2d_tfe_dmask0:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v3
|
|
; GFX6789-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v4
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_2d_tfe_dmask0:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v3, 0
|
|
; NOPRT-NEXT: image_load_mip v[2:3], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: v_mov_b32_e32 v0, v3
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_2d_tfe_dmask0:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, 0 ; encoding: [0x80,0x02,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v3 ; encoding: [0x03,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_2D unorm tfe ; encoding: [0x08,0x11,0x05,0xf0,0x00,0x03,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v4 ; encoding: [0x04,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.mip.2d.v4f32i32.i32(i32 0, i32 %s, i32 %t, i32 %mip, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
%vv = bitcast i32 %v.err to float
|
|
ret float %vv
|
|
}
|
|
|
|
define amdgpu_ps float @load_mip_2d_tfe_nouse(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_2d_tfe_nouse:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v3
|
|
; VERDE-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, v4
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_2d_tfe_nouse:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v3
|
|
; FIJI-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, v4
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_2d_tfe_nouse:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v3
|
|
; GFX6789-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v4
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_2d_tfe_nouse:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v3, 0
|
|
; NOPRT-NEXT: image_load_mip v[2:3], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: v_mov_b32_e32 v0, v3
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_2d_tfe_nouse:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, 0 ; encoding: [0x80,0x02,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v3 ; encoding: [0x03,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_2D unorm tfe ; encoding: [0x08,0x11,0x05,0xf0,0x00,0x03,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v4 ; encoding: [0x04,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.mip.2d.v4f32i32.i32(i32 15, i32 %s, i32 %t, i32 %mip, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
%vv = bitcast i32 %v.err to float
|
|
ret float %vv
|
|
}
|
|
|
|
define amdgpu_ps float @load_mip_2d_tfe_nouse_V2(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_2d_tfe_nouse_V2:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v3
|
|
; VERDE-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, v4
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_2d_tfe_nouse_V2:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v3
|
|
; FIJI-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, v4
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_2d_tfe_nouse_V2:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v3
|
|
; GFX6789-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v4
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_2d_tfe_nouse_V2:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v3, 0
|
|
; NOPRT-NEXT: image_load_mip v[2:3], v[0:2], s[0:7] dmask:0x1 unorm tfe
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: v_mov_b32_e32 v0, v3
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_2d_tfe_nouse_V2:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, 0 ; encoding: [0x80,0x02,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v3 ; encoding: [0x03,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_2D unorm tfe ; encoding: [0x08,0x11,0x05,0xf0,0x00,0x03,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v4 ; encoding: [0x04,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<2 x float>,i32} @llvm.amdgcn.image.load.mip.2d.v2f32i32.i32(i32 6, i32 %s, i32 %t, i32 %mip, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.err = extractvalue {<2 x float>, i32} %v, 1
|
|
%vv = bitcast i32 %v.err to float
|
|
ret float %vv
|
|
}
|
|
|
|
define amdgpu_ps float @load_mip_2d_tfe_nouse_V1(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_2d_tfe_nouse_V1:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v3
|
|
; VERDE-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x2 unorm tfe
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, v4
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_2d_tfe_nouse_V1:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v3
|
|
; FIJI-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x2 unorm tfe
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, v4
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_2d_tfe_nouse_V1:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v3
|
|
; GFX6789-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x2 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v4
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_2d_tfe_nouse_V1:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v3, 0
|
|
; NOPRT-NEXT: image_load_mip v[2:3], v[0:2], s[0:7] dmask:0x2 unorm tfe
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: v_mov_b32_e32 v0, v3
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_2d_tfe_nouse_V1:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, 0 ; encoding: [0x80,0x02,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v3 ; encoding: [0x03,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: image_load_mip v[3:4], v[0:2], s[0:7] dmask:0x2 dim:SQ_RSRC_IMG_2D unorm tfe ; encoding: [0x08,0x12,0x05,0xf0,0x00,0x03,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v4 ; encoding: [0x04,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {float, i32} @llvm.amdgcn.image.load.mip.2d.f32i32.i32(i32 2, i32 %s, i32 %t, i32 %mip, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.err = extractvalue {float, i32} %v, 1
|
|
%vv = bitcast i32 %v.err to float
|
|
ret float %vv
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1d_tfe_V4_dmask3(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s) {
|
|
; VERDE-LABEL: load_1d_tfe_V4_dmask3:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v4, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: image_load v[0:3], v4, s[0:7] dmask:0x7 unorm tfe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v3, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_tfe_V4_dmask3:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v4, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: image_load v[0:3], v4, s[0:7] dmask:0x7 unorm tfe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v3, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_tfe_V4_dmask3:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v5
|
|
; GFX6789-NEXT: v_mov_b32_e32 v7, v5
|
|
; GFX6789-NEXT: v_mov_b32_e32 v8, v5
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v5
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v6
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v7
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v8
|
|
; GFX6789-NEXT: image_load v[0:3], v4, s[0:7] dmask:0x7 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v5, v3, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_tfe_V4_dmask3:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v3, 0
|
|
; NOPRT-NEXT: image_load v[0:3], v0, s[0:7] dmask:0x7 unorm tfe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v4, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v4, v3, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_tfe_V4_dmask3:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, 0 ; encoding: [0x80,0x02,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v0 ; encoding: [0x00,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v5 ; encoding: [0x05,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v7, v5 ; encoding: [0x05,0x03,0x0e,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v8, v5 ; encoding: [0x05,0x03,0x10,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v5 ; encoding: [0x05,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v6 ; encoding: [0x06,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v7 ; encoding: [0x07,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v8 ; encoding: [0x08,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: image_load v[0:3], v4, s[0:7] dmask:0x7 dim:SQ_RSRC_IMG_1D unorm tfe ; encoding: [0x00,0x17,0x01,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v5, v3, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x05,0x03,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.1d.v4f32i32.i32(i32 7, i32 %s, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1d_tfe_V4_dmask2(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s) {
|
|
; VERDE-LABEL: load_1d_tfe_V4_dmask2:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: image_load v[0:2], v3, s[0:7] dmask:0x6 unorm tfe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v2, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_tfe_V4_dmask2:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: image_load v[0:2], v3, s[0:7] dmask:0x6 unorm tfe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v2, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_tfe_V4_dmask2:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v5, v4
|
|
; GFX6789-NEXT: v_mov_b32_e32 v6, v4
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v4
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v5
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v6
|
|
; GFX6789-NEXT: image_load v[0:2], v3, s[0:7] dmask:0x6 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v4, v2, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_tfe_V4_dmask2:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v2, 0
|
|
; NOPRT-NEXT: image_load v[0:2], v0, s[0:7] dmask:0x6 unorm tfe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v3, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v3, v2, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_tfe_V4_dmask2:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, 0 ; encoding: [0x80,0x02,0x08,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, v0 ; encoding: [0x00,0x03,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v5, v4 ; encoding: [0x04,0x03,0x0a,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v6, v4 ; encoding: [0x04,0x03,0x0c,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v4 ; encoding: [0x04,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v5 ; encoding: [0x05,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v6 ; encoding: [0x06,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: image_load v[0:2], v3, s[0:7] dmask:0x6 dim:SQ_RSRC_IMG_1D unorm tfe ; encoding: [0x00,0x16,0x01,0xf0,0x03,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v4, v2, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x04,0x02,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.1d.v4f32i32.i32(i32 6, i32 %s, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1d_tfe_V4_dmask1(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s) {
|
|
; VERDE-LABEL: load_1d_tfe_V4_dmask1:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: image_load v[0:1], v2, s[0:7] dmask:0x8 unorm tfe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v1, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_tfe_V4_dmask1:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: image_load v[0:1], v2, s[0:7] dmask:0x8 unorm tfe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v1, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_tfe_V4_dmask1:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v3
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v3
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v4
|
|
; GFX6789-NEXT: image_load v[0:1], v2, s[0:7] dmask:0x8 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v3, v1, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_tfe_V4_dmask1:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v1, 0
|
|
; NOPRT-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x8 unorm tfe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v2, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v2, v1, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_tfe_V4_dmask1:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, 0 ; encoding: [0x80,0x02,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v0 ; encoding: [0x00,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v3 ; encoding: [0x03,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v3 ; encoding: [0x03,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v4 ; encoding: [0x04,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: image_load v[0:1], v2, s[0:7] dmask:0x8 dim:SQ_RSRC_IMG_1D unorm tfe ; encoding: [0x00,0x18,0x01,0xf0,0x02,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v3, v1, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x03,0x01,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<4 x float>,i32} @llvm.amdgcn.image.load.1d.v4f32i32.i32(i32 8, i32 %s, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<4 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<4 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <4 x float> %v.vec
|
|
}
|
|
|
|
define amdgpu_ps <2 x float> @load_1d_tfe_V2_dmask1(<8 x i32> inreg %rsrc, i32 addrspace(1)* inreg %out, i32 %s) {
|
|
; VERDE-LABEL: load_1d_tfe_V2_dmask1:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: v_mov_b32_e32 v2, v0
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, 0
|
|
; VERDE-NEXT: v_mov_b32_e32 v1, v0
|
|
; VERDE-NEXT: image_load v[0:1], v2, s[0:7] dmask:0x8 unorm tfe
|
|
; VERDE-NEXT: s_mov_b32 s11, 0xf000
|
|
; VERDE-NEXT: s_mov_b32 s10, -1
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: buffer_store_dword v1, off, s[8:11], 0
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0) expcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_tfe_V2_dmask1:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: v_mov_b32_e32 v2, v0
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, 0
|
|
; FIJI-NEXT: v_mov_b32_e32 v1, v0
|
|
; FIJI-NEXT: image_load v[0:1], v2, s[0:7] dmask:0x8 unorm tfe
|
|
; FIJI-NEXT: s_mov_b32 s11, 0xf000
|
|
; FIJI-NEXT: s_mov_b32 s10, -1
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: buffer_store_dword v1, off, s[8:11], 0
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_tfe_V2_dmask1:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, 0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v2, v0
|
|
; GFX6789-NEXT: v_mov_b32_e32 v4, v3
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v3
|
|
; GFX6789-NEXT: v_mov_b32_e32 v1, v4
|
|
; GFX6789-NEXT: image_load v[0:1], v2, s[0:7] dmask:0x8 unorm tfe
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: global_store_dword v3, v1, s[8:9]
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_tfe_V2_dmask1:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: v_mov_b32_e32 v1, 0
|
|
; NOPRT-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x8 unorm tfe
|
|
; NOPRT-NEXT: v_mov_b32_e32 v2, 0
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: global_store_dword v2, v1, s[8:9]
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_tfe_V2_dmask1:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: v_mov_b32_e32 v3, 0 ; encoding: [0x80,0x02,0x06,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, v0 ; encoding: [0x00,0x03,0x04,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v4, v3 ; encoding: [0x03,0x03,0x08,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v3 ; encoding: [0x03,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: v_mov_b32_e32 v1, v4 ; encoding: [0x04,0x03,0x02,0x7e]
|
|
; GFX10-NEXT: image_load v[0:1], v2, s[0:7] dmask:0x8 dim:SQ_RSRC_IMG_1D unorm tfe ; encoding: [0x00,0x18,0x01,0xf0,0x02,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: global_store_dword v3, v1, s[8:9] ; encoding: [0x00,0x80,0x70,0xdc,0x03,0x01,0x08,0x00]
|
|
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0 ; encoding: [0x00,0x00,0xfd,0xbb]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call {<2 x float>,i32} @llvm.amdgcn.image.load.1d.v2f32i32.i32(i32 8, i32 %s, <8 x i32> %rsrc, i32 1, i32 0)
|
|
%v.vec = extractvalue {<2 x float>, i32} %v, 0
|
|
%v.err = extractvalue {<2 x float>, i32} %v, 1
|
|
store i32 %v.err, i32 addrspace(1)* %out, align 4
|
|
ret <2 x float> %v.vec
|
|
}
|
|
|
|
|
|
define amdgpu_ps <4 x float> @load_mip_3d(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %r, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_3d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_3d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_3d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_3d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_3d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm ; encoding: [0x10,0x1f,0x04,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.mip.3d.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %r, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_mip_cube(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %slice, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_cube:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_cube:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_cube:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_cube:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_cube:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm ; encoding: [0x18,0x1f,0x04,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.mip.cube.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %slice, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_mip_1darray(<8 x i32> inreg %rsrc, i32 %s, i32 %slice, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_1darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_1darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_1darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_1darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_1darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load_mip v[0:3], v[0:2], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm ; encoding: [0x20,0x1f,0x04,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.mip.1darray.v4f32.i32(i32 15, i32 %s, i32 %slice, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_mip_2darray(<8 x i32> inreg %rsrc, i32 %s, i32 %t, i32 %slice, i32 %mip) {
|
|
; VERDE-LABEL: load_mip_2darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_mip_2darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_mip_2darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_mip_2darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_mip_2darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load_mip v[0:3], v[0:3], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm ; encoding: [0x28,0x1f,0x04,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.mip.2darray.v4f32.i32(i32 15, i32 %s, i32 %t, i32 %slice, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps void @store_1d(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s) {
|
|
; VERDE-LABEL: store_1d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_1d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_1d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_1d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_1d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.1d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_2d(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t) {
|
|
; VERDE-LABEL: store_2d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_2d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_2d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_2d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_2d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm ; encoding: [0x08,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.2d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_3d(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %r) {
|
|
; VERDE-LABEL: store_3d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_3d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_3d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_3d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_3d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm ; encoding: [0x10,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.3d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %r, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_cube(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %slice) {
|
|
; VERDE-LABEL: store_cube:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_cube:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_cube:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_cube:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_cube:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm ; encoding: [0x18,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.cube.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %slice, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_1darray(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %slice) {
|
|
; VERDE-LABEL: store_1darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_1darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_1darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_1darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_1darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm ; encoding: [0x20,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.1darray.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %slice, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_2darray(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %slice) {
|
|
; VERDE-LABEL: store_2darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_2darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_2darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_2darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_2darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm ; encoding: [0x28,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.2darray.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %slice, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_2dmsaa(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %fragid) {
|
|
; VERDE-LABEL: store_2dmsaa:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_2dmsaa:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_2dmsaa:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_2dmsaa:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_2dmsaa:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v[4:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm ; encoding: [0x30,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.2dmsaa.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_2darraymsaa(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %slice, i32 %fragid) {
|
|
; VERDE-LABEL: store_2darraymsaa:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_2darraymsaa:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_2darraymsaa:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_2darraymsaa:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_2darraymsaa:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v[4:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm ; encoding: [0x38,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.2darraymsaa.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %slice, i32 %fragid, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_mip_1d(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %mip) {
|
|
; VERDE-LABEL: store_mip_1d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_mip_1d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_mip_1d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_mip_1d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_mip_1d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store_mip v[0:3], v[4:5], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1f,0x24,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.mip.1d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_mip_2d(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %mip) {
|
|
; VERDE-LABEL: store_mip_2d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_mip_2d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_mip_2d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_mip_2d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_mip_2d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm ; encoding: [0x08,0x1f,0x24,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.mip.2d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_mip_3d(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %r, i32 %mip) {
|
|
; VERDE-LABEL: store_mip_3d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_mip_3d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_mip_3d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_mip_3d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_mip_3d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm ; encoding: [0x10,0x1f,0x24,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.mip.3d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %r, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_mip_cube(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %slice, i32 %mip) {
|
|
; VERDE-LABEL: store_mip_cube:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_mip_cube:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_mip_cube:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_mip_cube:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_mip_cube:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm ; encoding: [0x18,0x1f,0x24,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.mip.cube.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %slice, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_mip_1darray(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %slice, i32 %mip) {
|
|
; VERDE-LABEL: store_mip_1darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_mip_1darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_mip_1darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_mip_1darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_mip_1darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store_mip v[0:3], v[4:6], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm ; encoding: [0x20,0x1f,0x24,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.mip.1darray.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %slice, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_mip_2darray(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s, i32 %t, i32 %slice, i32 %mip) {
|
|
; VERDE-LABEL: store_mip_2darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_mip_2darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_mip_2darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_mip_2darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_mip_2darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store_mip v[0:3], v[4:7], s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm ; encoding: [0x28,0x1f,0x24,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.mip.2darray.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, i32 %t, i32 %slice, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_1d(<8 x i32> inreg %rsrc, i32 %mip) {
|
|
; VERDE-LABEL: getresinfo_1d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_1d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_1d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_1d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_1d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1f,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.getresinfo.1d.v4f32.i32(i32 15, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_2d(<8 x i32> inreg %rsrc, i32 %mip) {
|
|
; VERDE-LABEL: getresinfo_2d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_2d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_2d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_2d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_2d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D unorm ; encoding: [0x08,0x1f,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.getresinfo.2d.v4f32.i32(i32 15, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_3d(<8 x i32> inreg %rsrc, i32 %mip) {
|
|
; VERDE-LABEL: getresinfo_3d:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_3d:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_3d:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_3d:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_3d:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_3D unorm ; encoding: [0x10,0x1f,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.getresinfo.3d.v4f32.i32(i32 15, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_cube(<8 x i32> inreg %rsrc, i32 %mip) {
|
|
; VERDE-LABEL: getresinfo_cube:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_cube:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_cube:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_cube:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_cube:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_CUBE unorm ; encoding: [0x18,0x1f,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.getresinfo.cube.v4f32.i32(i32 15, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_1darray(<8 x i32> inreg %rsrc, i32 %mip) {
|
|
; VERDE-LABEL: getresinfo_1darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_1darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_1darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_1darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_1darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D_ARRAY unorm ; encoding: [0x20,0x1f,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.getresinfo.1darray.v4f32.i32(i32 15, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_2darray(<8 x i32> inreg %rsrc, i32 %mip) {
|
|
; VERDE-LABEL: getresinfo_2darray:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_2darray:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_2darray:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_2darray:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_2darray:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_ARRAY unorm ; encoding: [0x28,0x1f,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.getresinfo.2darray.v4f32.i32(i32 15, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_2dmsaa(<8 x i32> inreg %rsrc, i32 %mip) {
|
|
; VERDE-LABEL: getresinfo_2dmsaa:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_2dmsaa:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_2dmsaa:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_2dmsaa:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_2dmsaa:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA unorm ; encoding: [0x30,0x1f,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.getresinfo.2dmsaa.v4f32.i32(i32 15, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_2darraymsaa(<8 x i32> inreg %rsrc, i32 %mip) {
|
|
; VERDE-LABEL: getresinfo_2darraymsaa:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_2darraymsaa:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_2darraymsaa:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_2darraymsaa:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf unorm da
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_2darraymsaa:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_2D_MSAA_ARRAY unorm ; encoding: [0x38,0x1f,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.getresinfo.2darraymsaa.v4f32.i32(i32 15, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps float @load_1d_V1(<8 x i32> inreg %rsrc, i32 %s) {
|
|
; VERDE-LABEL: load_1d_V1:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v0, v0, s[0:7] dmask:0x8 unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_V1:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v0, v0, s[0:7] dmask:0x8 unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_V1:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v0, v0, s[0:7] dmask:0x8 unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_V1:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v0, v0, s[0:7] dmask:0x8 unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_V1:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v0, v0, s[0:7] dmask:0x8 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x18,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call float @llvm.amdgcn.image.load.1d.f32.i32(i32 8, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret float %v
|
|
}
|
|
|
|
define amdgpu_ps <2 x float> @load_1d_V2(<8 x i32> inreg %rsrc, i32 %s) {
|
|
; VERDE-LABEL: load_1d_V2:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x9 unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_V2:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x9 unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_V2:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x9 unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_V2:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x9 unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_V2:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:1], v0, s[0:7] dmask:0x9 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x19,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <2 x float> @llvm.amdgcn.image.load.1d.v2f32.i32(i32 9, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <2 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps void @store_1d_V1(<8 x i32> inreg %rsrc, float %vdata, i32 %s) {
|
|
; VERDE-LABEL: store_1d_V1:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v0, v1, s[0:7] dmask:0x2 unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_1d_V1:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v0, v1, s[0:7] dmask:0x2 unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_1d_V1:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v0, v1, s[0:7] dmask:0x2 unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_1d_V1:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v0, v1, s[0:7] dmask:0x2 unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_1d_V1:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v0, v1, s[0:7] dmask:0x2 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x12,0x20,0xf0,0x01,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.1d.f32.i32(float %vdata, i32 2, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_1d_V2(<8 x i32> inreg %rsrc, <2 x float> %vdata, i32 %s) {
|
|
; VERDE-LABEL: store_1d_V2:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:1], v2, s[0:7] dmask:0xc unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_1d_V2:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:1], v2, s[0:7] dmask:0xc unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_1d_V2:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:1], v2, s[0:7] dmask:0xc unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_1d_V2:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:1], v2, s[0:7] dmask:0xc unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_1d_V2:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:1], v2, s[0:7] dmask:0xc dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1c,0x20,0xf0,0x02,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.1d.v2f32.i32(<2 x float> %vdata, i32 12, i32 %s, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1d_glc(<8 x i32> inreg %rsrc, i32 %s) {
|
|
; VERDE-LABEL: load_1d_glc:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_glc:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_glc:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_glc:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_glc:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc ; encoding: [0x00,0x3f,0x00,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i32(i32 15, i32 %s, <8 x i32> %rsrc, i32 0, i32 1)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1d_slc(<8 x i32> inreg %rsrc, i32 %s) {
|
|
; VERDE-LABEL: load_1d_slc:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm slc
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_slc:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm slc
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_slc:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm slc
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_slc:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm slc
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_slc:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm slc ; encoding: [0x00,0x1f,0x00,0xf2,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i32(i32 15, i32 %s, <8 x i32> %rsrc, i32 0, i32 2)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @load_1d_glc_slc(<8 x i32> inreg %rsrc, i32 %s) {
|
|
; VERDE-LABEL: load_1d_glc_slc:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc slc
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: load_1d_glc_slc:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc slc
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: load_1d_glc_slc:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc slc
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: load_1d_glc_slc:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf unorm glc slc
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: load_1d_glc_slc:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_load v[0:3], v0, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc slc ; encoding: [0x00,0x3f,0x00,0xf2,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%v = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i32(i32 15, i32 %s, <8 x i32> %rsrc, i32 0, i32 3)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
define amdgpu_ps void @store_1d_glc(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s) {
|
|
; VERDE-LABEL: store_1d_glc:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_1d_glc:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_1d_glc:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_1d_glc:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_1d_glc:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc ; encoding: [0x00,0x3f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.1d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, <8 x i32> %rsrc, i32 0, i32 1)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_1d_slc(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s) {
|
|
; VERDE-LABEL: store_1d_slc:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm slc
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_1d_slc:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm slc
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_1d_slc:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm slc
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_1d_slc:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm slc
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_1d_slc:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm slc ; encoding: [0x00,0x1f,0x20,0xf2,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.1d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, <8 x i32> %rsrc, i32 0, i32 2)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps void @store_1d_glc_slc(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %s) {
|
|
; VERDE-LABEL: store_1d_glc_slc:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc slc
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: store_1d_glc_slc:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc slc
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: store_1d_glc_slc:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc slc
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: store_1d_glc_slc:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm glc slc
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: store_1d_glc_slc:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm glc slc ; encoding: [0x00,0x3f,0x20,0xf2,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.1d.v4f32.i32(<4 x float> %vdata, i32 15, i32 %s, <8 x i32> %rsrc, i32 0, i32 3)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps <3 x float> @getresinfo_dmask7(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %mip) {
|
|
; GFX6-LABEL: getresinfo_dmask7:
|
|
; GFX6: ; %bb.0: ; %main_body
|
|
; GFX6-NEXT: s_mov_b32 s0, s2
|
|
; GFX6-NEXT: s_mov_b32 s1, s3
|
|
; GFX6-NEXT: s_mov_b32 s2, s4
|
|
; GFX6-NEXT: s_mov_b32 s3, s5
|
|
; GFX6-NEXT: s_mov_b32 s4, s6
|
|
; GFX6-NEXT: s_mov_b32 s5, s7
|
|
; GFX6-NEXT: s_mov_b32 s6, s8
|
|
; GFX6-NEXT: s_mov_b32 s7, s9
|
|
; GFX6-NEXT: image_get_resinfo v[0:2], v0, s[0:7] dmask:0x7 unorm
|
|
; GFX6-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-LABEL: getresinfo_dmask7:
|
|
; GFX8: ; %bb.0: ; %main_body
|
|
; GFX8-NEXT: s_mov_b32 s0, s2
|
|
; GFX8-NEXT: s_mov_b32 s1, s3
|
|
; GFX8-NEXT: s_mov_b32 s2, s4
|
|
; GFX8-NEXT: s_mov_b32 s3, s5
|
|
; GFX8-NEXT: s_mov_b32 s4, s6
|
|
; GFX8-NEXT: s_mov_b32 s5, s7
|
|
; GFX8-NEXT: s_mov_b32 s6, s8
|
|
; GFX8-NEXT: s_mov_b32 s7, s9
|
|
; GFX8-NEXT: image_get_resinfo v[0:2], v0, s[0:7] dmask:0x7 unorm
|
|
; GFX8-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX8-NEXT: ; return to shader part epilog
|
|
;
|
|
; VERDE-LABEL: getresinfo_dmask7:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:2], v0, s[0:7] dmask:0x7 unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_dmask7:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:2], v0, s[0:7] dmask:0x7 unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_dmask7:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:2], v0, s[0:7] dmask:0x7 unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_dmask7:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:2], v0, s[0:7] dmask:0x7 unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_dmask7:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:2], v0, s[0:7] dmask:0x7 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x17,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%r = call <3 x float> @llvm.amdgcn.image.getresinfo.1d.v3f32.i32(i32 7, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <3 x float> %r
|
|
}
|
|
|
|
define amdgpu_ps <2 x float> @getresinfo_dmask3(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %mip) {
|
|
; GFX6-LABEL: getresinfo_dmask3:
|
|
; GFX6: ; %bb.0: ; %main_body
|
|
; GFX6-NEXT: s_mov_b32 s0, s2
|
|
; GFX6-NEXT: s_mov_b32 s1, s3
|
|
; GFX6-NEXT: s_mov_b32 s2, s4
|
|
; GFX6-NEXT: s_mov_b32 s3, s5
|
|
; GFX6-NEXT: s_mov_b32 s4, s6
|
|
; GFX6-NEXT: s_mov_b32 s5, s7
|
|
; GFX6-NEXT: s_mov_b32 s6, s8
|
|
; GFX6-NEXT: s_mov_b32 s7, s9
|
|
; GFX6-NEXT: image_get_resinfo v[0:1], v0, s[0:7] dmask:0x3 unorm
|
|
; GFX6-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-LABEL: getresinfo_dmask3:
|
|
; GFX8: ; %bb.0: ; %main_body
|
|
; GFX8-NEXT: s_mov_b32 s0, s2
|
|
; GFX8-NEXT: s_mov_b32 s1, s3
|
|
; GFX8-NEXT: s_mov_b32 s2, s4
|
|
; GFX8-NEXT: s_mov_b32 s3, s5
|
|
; GFX8-NEXT: s_mov_b32 s4, s6
|
|
; GFX8-NEXT: s_mov_b32 s5, s7
|
|
; GFX8-NEXT: s_mov_b32 s6, s8
|
|
; GFX8-NEXT: s_mov_b32 s7, s9
|
|
; GFX8-NEXT: image_get_resinfo v[0:1], v0, s[0:7] dmask:0x3 unorm
|
|
; GFX8-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX8-NEXT: ; return to shader part epilog
|
|
;
|
|
; VERDE-LABEL: getresinfo_dmask3:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v[0:1], v0, s[0:7] dmask:0x3 unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_dmask3:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v[0:1], v0, s[0:7] dmask:0x3 unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_dmask3:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v[0:1], v0, s[0:7] dmask:0x3 unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_dmask3:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v[0:1], v0, s[0:7] dmask:0x3 unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_dmask3:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v[0:1], v0, s[0:7] dmask:0x3 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x13,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%r = call <2 x float> @llvm.amdgcn.image.getresinfo.1d.v2f32.i32(i32 3, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <2 x float> %r
|
|
}
|
|
|
|
define amdgpu_ps float @getresinfo_dmask1(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %mip) {
|
|
; GFX6-LABEL: getresinfo_dmask1:
|
|
; GFX6: ; %bb.0: ; %main_body
|
|
; GFX6-NEXT: s_mov_b32 s0, s2
|
|
; GFX6-NEXT: s_mov_b32 s1, s3
|
|
; GFX6-NEXT: s_mov_b32 s2, s4
|
|
; GFX6-NEXT: s_mov_b32 s3, s5
|
|
; GFX6-NEXT: s_mov_b32 s4, s6
|
|
; GFX6-NEXT: s_mov_b32 s5, s7
|
|
; GFX6-NEXT: s_mov_b32 s6, s8
|
|
; GFX6-NEXT: s_mov_b32 s7, s9
|
|
; GFX6-NEXT: image_get_resinfo v0, v0, s[0:7] dmask:0x1 unorm
|
|
; GFX6-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX8-LABEL: getresinfo_dmask1:
|
|
; GFX8: ; %bb.0: ; %main_body
|
|
; GFX8-NEXT: s_mov_b32 s0, s2
|
|
; GFX8-NEXT: s_mov_b32 s1, s3
|
|
; GFX8-NEXT: s_mov_b32 s2, s4
|
|
; GFX8-NEXT: s_mov_b32 s3, s5
|
|
; GFX8-NEXT: s_mov_b32 s4, s6
|
|
; GFX8-NEXT: s_mov_b32 s5, s7
|
|
; GFX8-NEXT: s_mov_b32 s6, s8
|
|
; GFX8-NEXT: s_mov_b32 s7, s9
|
|
; GFX8-NEXT: image_get_resinfo v0, v0, s[0:7] dmask:0x1 unorm
|
|
; GFX8-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX8-NEXT: ; return to shader part epilog
|
|
;
|
|
; VERDE-LABEL: getresinfo_dmask1:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_get_resinfo v0, v0, s[0:7] dmask:0x1 unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_dmask1:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_get_resinfo v0, v0, s[0:7] dmask:0x1 unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_dmask1:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_get_resinfo v0, v0, s[0:7] dmask:0x1 unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_dmask1:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_get_resinfo v0, v0, s[0:7] dmask:0x1 unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_dmask1:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_get_resinfo v0, v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x11,0x38,0xf0,0x00,0x00,0x00,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%r = call float @llvm.amdgcn.image.getresinfo.1d.f32.i32(i32 1, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret float %r
|
|
}
|
|
|
|
define amdgpu_ps <4 x float> @getresinfo_dmask0(<8 x i32> inreg %rsrc, <4 x float> %vdata, i32 %mip) #0 {
|
|
; VERDE-LABEL: getresinfo_dmask0:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: getresinfo_dmask0:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: getresinfo_dmask0:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: getresinfo_dmask0:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: getresinfo_dmask0:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
main_body:
|
|
%r = call <4 x float> @llvm.amdgcn.image.getresinfo.1d.v4f32.i32(i32 0, i32 %mip, <8 x i32> %rsrc, i32 0, i32 0)
|
|
ret <4 x float> %r
|
|
}
|
|
|
|
;
|
|
define amdgpu_ps void @image_store_wait(<8 x i32> inreg %arg, <8 x i32> inreg %arg1, <8 x i32> inreg %arg2, <4 x float> %arg3, i32 %arg4) #0 {
|
|
; VERDE-LABEL: image_store_wait:
|
|
; VERDE: ; %bb.0: ; %main_body
|
|
; VERDE-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt expcnt(0)
|
|
; VERDE-NEXT: image_load v[0:3], v4, s[8:15] dmask:0xf unorm
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: image_store v[0:3], v4, s[16:23] dmask:0xf unorm
|
|
; VERDE-NEXT: s_endpgm
|
|
;
|
|
; FIJI-LABEL: image_store_wait:
|
|
; FIJI: ; %bb.0: ; %main_body
|
|
; FIJI-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm
|
|
; FIJI-NEXT: image_load v[0:3], v4, s[8:15] dmask:0xf unorm
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: image_store v[0:3], v4, s[16:23] dmask:0xf unorm
|
|
; FIJI-NEXT: s_endpgm
|
|
;
|
|
; GFX6789-LABEL: image_store_wait:
|
|
; GFX6789: ; %bb.0: ; %main_body
|
|
; GFX6789-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm
|
|
; GFX6789-NEXT: image_load v[0:3], v4, s[8:15] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: image_store v[0:3], v4, s[16:23] dmask:0xf unorm
|
|
; GFX6789-NEXT: s_endpgm
|
|
;
|
|
; NOPRT-LABEL: image_store_wait:
|
|
; NOPRT: ; %bb.0: ; %main_body
|
|
; NOPRT-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf unorm
|
|
; NOPRT-NEXT: image_load v[0:3], v4, s[8:15] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: image_store v[0:3], v4, s[16:23] dmask:0xf unorm
|
|
; NOPRT-NEXT: s_endpgm
|
|
;
|
|
; GFX10-LABEL: image_store_wait:
|
|
; GFX10: ; %bb.0: ; %main_body
|
|
; GFX10-NEXT: image_store v[0:3], v4, s[0:7] dmask:0xf dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1f,0x20,0xf0,0x04,0x00,0x00,0x00]
|
|
; GFX10-NEXT: image_load v[0:3], v4, s[8:15] dmask:0xf dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1f,0x00,0xf0,0x04,0x00,0x02,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: image_store v[0:3], v4, s[16:23] dmask:0xf dim:SQ_RSRC_IMG_1D unorm ; encoding: [0x00,0x1f,0x20,0xf0,0x04,0x00,0x04,0x00]
|
|
; GFX10-NEXT: s_endpgm ; encoding: [0x00,0x00,0x81,0xbf]
|
|
main_body:
|
|
call void @llvm.amdgcn.image.store.1d.v4f32.i32(<4 x float> %arg3, i32 15, i32 %arg4, <8 x i32> %arg, i32 0, i32 0)
|
|
%data = call <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i32(i32 15, i32 %arg4, <8 x i32> %arg1, i32 0, i32 0)
|
|
call void @llvm.amdgcn.image.store.1d.v4f32.i32(<4 x float> %data, i32 15, i32 %arg4, <8 x i32> %arg2, i32 0, i32 0)
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_ps float @image_load_mmo(<8 x i32> inreg %rsrc, float addrspace(3)* %lds, <2 x i32> %c) #0 {
|
|
; VERDE-LABEL: image_load_mmo:
|
|
; VERDE: ; %bb.0:
|
|
; VERDE-NEXT: image_load v1, v[1:2], s[0:7] dmask:0x1 unorm
|
|
; VERDE-NEXT: v_mov_b32_e32 v3, 0
|
|
; VERDE-NEXT: s_mov_b32 m0, -1
|
|
; VERDE-NEXT: ds_write_b32 v0, v3
|
|
; VERDE-NEXT: v_add_i32_e32 v0, vcc, 16, v0
|
|
; VERDE-NEXT: ds_write_b32 v0, v3
|
|
; VERDE-NEXT: s_waitcnt vmcnt(0)
|
|
; VERDE-NEXT: v_mov_b32_e32 v0, v1
|
|
; VERDE-NEXT: s_waitcnt lgkmcnt(0)
|
|
; VERDE-NEXT: ; return to shader part epilog
|
|
;
|
|
; FIJI-LABEL: image_load_mmo:
|
|
; FIJI: ; %bb.0:
|
|
; FIJI-NEXT: image_load v1, v[1:2], s[0:7] dmask:0x1 unorm
|
|
; FIJI-NEXT: v_mov_b32_e32 v3, 0
|
|
; FIJI-NEXT: s_mov_b32 m0, -1
|
|
; FIJI-NEXT: ds_write2_b32 v0, v3, v3 offset1:4
|
|
; FIJI-NEXT: s_waitcnt vmcnt(0)
|
|
; FIJI-NEXT: v_mov_b32_e32 v0, v1
|
|
; FIJI-NEXT: s_waitcnt lgkmcnt(0)
|
|
; FIJI-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX6789-LABEL: image_load_mmo:
|
|
; GFX6789: ; %bb.0:
|
|
; GFX6789-NEXT: image_load v1, v[1:2], s[0:7] dmask:0x1 unorm
|
|
; GFX6789-NEXT: v_mov_b32_e32 v3, 0
|
|
; GFX6789-NEXT: ds_write2_b32 v0, v3, v3 offset1:4
|
|
; GFX6789-NEXT: s_waitcnt vmcnt(0)
|
|
; GFX6789-NEXT: v_mov_b32_e32 v0, v1
|
|
; GFX6789-NEXT: s_waitcnt lgkmcnt(0)
|
|
; GFX6789-NEXT: ; return to shader part epilog
|
|
;
|
|
; NOPRT-LABEL: image_load_mmo:
|
|
; NOPRT: ; %bb.0:
|
|
; NOPRT-NEXT: image_load v1, v[1:2], s[0:7] dmask:0x1 unorm
|
|
; NOPRT-NEXT: v_mov_b32_e32 v3, 0
|
|
; NOPRT-NEXT: ds_write2_b32 v0, v3, v3 offset1:4
|
|
; NOPRT-NEXT: s_waitcnt vmcnt(0)
|
|
; NOPRT-NEXT: v_mov_b32_e32 v0, v1
|
|
; NOPRT-NEXT: s_waitcnt lgkmcnt(0)
|
|
; NOPRT-NEXT: ; return to shader part epilog
|
|
;
|
|
; GFX10-LABEL: image_load_mmo:
|
|
; GFX10: ; %bb.0:
|
|
; GFX10-NEXT: image_load v1, v[1:2], s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_2D unorm ; encoding: [0x08,0x11,0x00,0xf0,0x01,0x01,0x00,0x00]
|
|
; GFX10-NEXT: v_mov_b32_e32 v2, 0 ; encoding: [0x80,0x02,0x04,0x7e]
|
|
; GFX10-NEXT: ds_write2_b32 v0, v2, v2 offset1:4 ; encoding: [0x00,0x04,0x38,0xd8,0x00,0x02,0x02,0x00]
|
|
; GFX10-NEXT: s_waitcnt vmcnt(0) ; encoding: [0x70,0x3f,0x8c,0xbf]
|
|
; GFX10-NEXT: v_mov_b32_e32 v0, v1 ; encoding: [0x01,0x03,0x00,0x7e]
|
|
; GFX10-NEXT: s_waitcnt lgkmcnt(0) ; encoding: [0x7f,0xc0,0x8c,0xbf]
|
|
; GFX10-NEXT: ; return to shader part epilog
|
|
store float 0.000000e+00, float addrspace(3)* %lds
|
|
%c0 = extractelement <2 x i32> %c, i32 0
|
|
%c1 = extractelement <2 x i32> %c, i32 1
|
|
%tex = call float @llvm.amdgcn.image.load.2d.f32.i32(i32 1, i32 %c0, i32 %c1, <8 x i32> %rsrc, i32 0, i32 0)
|
|
%tmp2 = getelementptr float, float addrspace(3)* %lds, i32 4
|
|
store float 0.000000e+00, float addrspace(3)* %tmp2
|
|
ret float %tex
|
|
}
|
|
|
|
declare <4 x float> @llvm.amdgcn.image.load.1d.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {float,i32} @llvm.amdgcn.image.load.1d.f32i32.i32(i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<2 x float>,i32} @llvm.amdgcn.image.load.1d.v2f32i32.i32(i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.1d.v4f32i32.i32(i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.2d.v4f32.i32(i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.2d.v4f32i32.i32(i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.3d.v4f32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.3d.v4f32i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.cube.v4f32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.cube.v4f32i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.1darray.v4f32.i32(i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.1darray.v4f32i32.i32(i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.2darray.v4f32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.2darray.v4f32i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.2dmsaa.v4f32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.2dmsaa.v4f32i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.2darraymsaa.v4f32.i32(i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.2darraymsaa.v4f32i32.i32(i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
|
|
declare <4 x float> @llvm.amdgcn.image.load.mip.1d.v4f32.i32(i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.mip.2d.v4f32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.mip.1d.v4f32i32.i32(i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<4 x float>,i32} @llvm.amdgcn.image.load.mip.2d.v4f32i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {<2 x float>,i32} @llvm.amdgcn.image.load.mip.2d.v2f32i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare {float,i32} @llvm.amdgcn.image.load.mip.2d.f32i32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.mip.3d.v4f32.i32(i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.mip.cube.v4f32.i32(i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.mip.1darray.v4f32.i32(i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <4 x float> @llvm.amdgcn.image.load.mip.2darray.v4f32.i32(i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
|
|
declare void @llvm.amdgcn.image.store.1d.v4f32.i32(<4 x float>, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.2d.v4f32.i32(<4 x float>, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.3d.v4f32.i32(<4 x float>, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.cube.v4f32.i32(<4 x float>, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.1darray.v4f32.i32(<4 x float>, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.2darray.v4f32.i32(<4 x float>, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.2dmsaa.v4f32.i32(<4 x float>, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.2darraymsaa.v4f32.i32(<4 x float>, i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
|
|
declare void @llvm.amdgcn.image.store.mip.1d.v4f32.i32(<4 x float>, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.mip.2d.v4f32.i32(<4 x float>, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.mip.3d.v4f32.i32(<4 x float>, i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.mip.cube.v4f32.i32(<4 x float>, i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.mip.1darray.v4f32.i32(<4 x float>, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.mip.2darray.v4f32.i32(<4 x float>, i32, i32, i32, i32, i32, <8 x i32>, i32, i32) #0
|
|
|
|
declare <4 x float> @llvm.amdgcn.image.getresinfo.1d.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #2
|
|
declare <3 x float> @llvm.amdgcn.image.getresinfo.1d.v3f32.i32(i32 immarg, i32, <8 x i32>, i32 immarg, i32 immarg) #1
|
|
declare <2 x float> @llvm.amdgcn.image.getresinfo.1d.v2f32.i32(i32 immarg, i32, <8 x i32>, i32 immarg, i32 immarg) #1
|
|
declare float @llvm.amdgcn.image.getresinfo.1d.f32.i32(i32 immarg, i32, <8 x i32>, i32 immarg, i32 immarg) #1
|
|
declare <4 x float> @llvm.amdgcn.image.getresinfo.2d.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #2
|
|
declare <4 x float> @llvm.amdgcn.image.getresinfo.3d.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #2
|
|
declare <4 x float> @llvm.amdgcn.image.getresinfo.cube.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #2
|
|
declare <4 x float> @llvm.amdgcn.image.getresinfo.1darray.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #2
|
|
declare <4 x float> @llvm.amdgcn.image.getresinfo.2darray.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #2
|
|
declare <4 x float> @llvm.amdgcn.image.getresinfo.2dmsaa.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #2
|
|
declare <4 x float> @llvm.amdgcn.image.getresinfo.2darraymsaa.v4f32.i32(i32, i32, <8 x i32>, i32, i32) #2
|
|
|
|
declare float @llvm.amdgcn.image.load.1d.f32.i32(i32, i32, <8 x i32>, i32, i32) #1
|
|
declare float @llvm.amdgcn.image.load.2d.f32.i32(i32, i32, i32, <8 x i32>, i32, i32) #1
|
|
declare <2 x float> @llvm.amdgcn.image.load.1d.v2f32.i32(i32, i32, <8 x i32>, i32, i32) #1
|
|
declare void @llvm.amdgcn.image.store.1d.f32.i32(float, i32, i32, <8 x i32>, i32, i32) #0
|
|
declare void @llvm.amdgcn.image.store.1d.v2f32.i32(<2 x float>, i32, i32, <8 x i32>, i32, i32) #0
|
|
|
|
attributes #0 = { nounwind }
|
|
attributes #1 = { nounwind readonly }
|
|
attributes #2 = { nounwind readnone }
|