Files
TheRock/compiler/amd-llvm/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.exp.row.ll
T
derek.holloway 7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

158 lines
6.1 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -global-isel=0 -mtriple=amdgcn -mcpu=gfx1100 < %s | FileCheck %s -check-prefixes=GFX11
; RUN: llc -global-isel=1 -new-reg-bank-select -mtriple=amdgcn -mcpu=gfx1100 < %s | FileCheck %s -check-prefixes=GFX11
; RUN: llc -global-isel=0 -mtriple=amdgcn -mcpu=gfx1200 < %s | FileCheck %s -check-prefixes=GFX12
; RUN: llc -global-isel=1 -new-reg-bank-select -mtriple=amdgcn -mcpu=gfx1200 < %s | FileCheck %s -check-prefixes=GFX12
declare void @llvm.amdgcn.exp.row.i32(i32, i32, i32, i32, i32, i32, i1, i32)
declare void @llvm.amdgcn.exp.row.f32(i32, i32, float, float, float, float, i1, i32)
declare i32 @llvm.amdgcn.workitem.id.x()
define amdgpu_kernel void @undef_i32() #0 {
; GFX11-LABEL: undef_i32:
; GFX11: ; %bb.0:
; GFX11-NEXT: s_mov_b32 m0, 0
; GFX11-NEXT: exp pos0, off, off, off, off row_en
; GFX11-NEXT: exp pos1, off, off, off, off done row_en
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: undef_i32:
; GFX12: ; %bb.0:
; GFX12-NEXT: s_mov_b32 m0, 0
; GFX12-NEXT: export pos0, off, off, off, off row_en
; GFX12-NEXT: export pos1, off, off, off, off done row_en
; GFX12-NEXT: s_endpgm
call void @llvm.amdgcn.exp.row.i32(i32 12, i32 0, i32 poison, i32 poison, i32 poison, i32 poison, i1 false, i32 0)
call void @llvm.amdgcn.exp.row.i32(i32 13, i32 0, i32 poison, i32 poison, i32 poison, i32 poison, i1 true, i32 0)
ret void
}
define amdgpu_kernel void @undef_f32() #0 {
; GFX11-LABEL: undef_f32:
; GFX11: ; %bb.0:
; GFX11-NEXT: s_mov_b32 m0, 0
; GFX11-NEXT: exp pos0, off, off, off, off row_en
; GFX11-NEXT: exp pos1, off, off, off, off done row_en
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: undef_f32:
; GFX12: ; %bb.0:
; GFX12-NEXT: s_mov_b32 m0, 0
; GFX12-NEXT: export pos0, off, off, off, off row_en
; GFX12-NEXT: export pos1, off, off, off, off done row_en
; GFX12-NEXT: s_endpgm
call void @llvm.amdgcn.exp.row.f32(i32 12, i32 0, float poison, float poison, float poison, float poison, i1 false, i32 0)
call void @llvm.amdgcn.exp.row.f32(i32 13, i32 0, float poison, float poison, float poison, float poison, i1 true, i32 0)
ret void
}
define amdgpu_kernel void @zero_i32() #0 {
; GFX11-LABEL: zero_i32:
; GFX11: ; %bb.0:
; GFX11-NEXT: v_mov_b32_e32 v0, 0
; GFX11-NEXT: s_mov_b32 m0, 0
; GFX11-NEXT: exp pos0, v0, v0, v0, off row_en
; GFX11-NEXT: exp pos1, v0, v0, v0, off done row_en
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: zero_i32:
; GFX12: ; %bb.0:
; GFX12-NEXT: v_mov_b32_e32 v0, 0
; GFX12-NEXT: s_mov_b32 m0, 0
; GFX12-NEXT: export pos0, v0, v0, v0, off row_en
; GFX12-NEXT: export pos1, v0, v0, v0, off done row_en
; GFX12-NEXT: s_endpgm
call void @llvm.amdgcn.exp.row.i32(i32 12, i32 7, i32 0, i32 0, i32 0, i32 poison, i1 false, i32 0)
call void @llvm.amdgcn.exp.row.i32(i32 13, i32 7, i32 0, i32 0, i32 0, i32 poison, i1 true, i32 0)
ret void
}
define amdgpu_kernel void @one_f32() #0 {
; GFX11-LABEL: one_f32:
; GFX11: ; %bb.0:
; GFX11-NEXT: v_mov_b32_e32 v0, 1.0
; GFX11-NEXT: s_mov_b32 m0, 0
; GFX11-NEXT: exp pos0, v0, v0, v0, off row_en
; GFX11-NEXT: exp pos1, v0, v0, v0, off done row_en
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: one_f32:
; GFX12: ; %bb.0:
; GFX12-NEXT: v_mov_b32_e32 v0, 1.0
; GFX12-NEXT: s_mov_b32 m0, 0
; GFX12-NEXT: export pos0, v0, v0, v0, off row_en
; GFX12-NEXT: export pos1, v0, v0, v0, off done row_en
; GFX12-NEXT: s_endpgm
call void @llvm.amdgcn.exp.row.f32(i32 12, i32 7, float 1.0, float 1.0, float 1.0, float poison, i1 false, i32 0)
call void @llvm.amdgcn.exp.row.f32(i32 13, i32 7, float 1.0, float 1.0, float 1.0, float poison, i1 true, i32 0)
ret void
}
define amdgpu_kernel void @id_i32() #0 {
; GFX11-LABEL: id_i32:
; GFX11: ; %bb.0:
; GFX11-NEXT: v_and_b32_e32 v0, 0x3ff, v0
; GFX11-NEXT: s_mov_b32 m0, 0
; GFX11-NEXT: exp pos0, v0, off, off, off done row_en
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: id_i32:
; GFX12: ; %bb.0:
; GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v0
; GFX12-NEXT: s_mov_b32 m0, 0
; GFX12-NEXT: export pos0, v0, off, off, off done row_en
; GFX12-NEXT: s_endpgm
%id = call i32 @llvm.amdgcn.workitem.id.x()
call void @llvm.amdgcn.exp.row.i32(i32 12, i32 1, i32 %id, i32 poison, i32 poison, i32 poison, i1 true, i32 0)
ret void
}
define amdgpu_kernel void @id_arg_i32(i32 %row) #0 {
; GFX11-LABEL: id_arg_i32:
; GFX11: ; %bb.0:
; GFX11-NEXT: s_load_b32 s0, s[4:5], 0x24
; GFX11-NEXT: v_and_b32_e32 v0, 0x3ff, v0
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
; GFX11-NEXT: s_mov_b32 m0, s0
; GFX11-NEXT: exp pos0, v0, off, off, off done row_en
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: id_arg_i32:
; GFX12: ; %bb.0:
; GFX12-NEXT: s_load_b32 s0, s[4:5], 0x24
; GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: s_mov_b32 m0, s0
; GFX12-NEXT: export pos0, v0, off, off, off done row_en
; GFX12-NEXT: s_endpgm
%id = call i32 @llvm.amdgcn.workitem.id.x()
call void @llvm.amdgcn.exp.row.i32(i32 12, i32 1, i32 %id, i32 poison, i32 poison, i32 poison, i1 true, i32 %row)
ret void
}
; Divergent row number just causes a readfirstlane for now.
define amdgpu_kernel void @id_row_i32() #0 {
; GFX11-LABEL: id_row_i32:
; GFX11: ; %bb.0:
; GFX11-NEXT: v_and_b32_e32 v0, 0x3ff, v0
; GFX11-NEXT: s_delay_alu instid0(VALU_DEP_1)
; GFX11-NEXT: v_readfirstlane_b32 s0, v0
; GFX11-NEXT: v_mov_b32_e32 v0, 0x63
; GFX11-NEXT: s_mov_b32 m0, s0
; GFX11-NEXT: exp pos0, v0, off, off, off done row_en
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: id_row_i32:
; GFX12: ; %bb.0:
; GFX12-NEXT: v_and_b32_e32 v0, 0x3ff, v0
; GFX12-NEXT: s_delay_alu instid0(VALU_DEP_1)
; GFX12-NEXT: v_readfirstlane_b32 s0, v0
; GFX12-NEXT: v_mov_b32_e32 v0, 0x63
; GFX12-NEXT: s_mov_b32 m0, s0
; GFX12-NEXT: export pos0, v0, off, off, off done row_en
; GFX12-NEXT: s_endpgm
%id = call i32 @llvm.amdgcn.workitem.id.x()
call void @llvm.amdgcn.exp.row.i32(i32 12, i32 1, i32 99, i32 poison, i32 poison, i32 poison, i1 true, i32 %id)
ret void
}