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
76 lines
5.4 KiB
LLVM
76 lines
5.4 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
|
|
; RUN: llc -mtriple=amdgcn -mcpu=gfx90a -stop-after=amdgpu-isel < %s | FileCheck -check-prefix=GFX90A_GFX942 %s
|
|
; RUN: llc -mtriple=amdgcn -mcpu=gfx942 -stop-after=amdgpu-isel < %s | FileCheck -check-prefix=GFX90A_GFX942 %s
|
|
; RUN: llc -mtriple=amdgcn -mcpu=gfx90a -enable-new-pm -stop-after=amdgpu-isel < %s | FileCheck -check-prefix=GFX90A_GFX942 %s
|
|
; RUN: llc -mtriple=amdgcn -mcpu=gfx942 -enable-new-pm -stop-after=amdgpu-isel < %s | FileCheck -check-prefix=GFX90A_GFX942 %s
|
|
|
|
define amdgpu_ps <2 x half> @global_atomic_fadd_v2f16_rtn(ptr addrspace(1) %ptr, <2 x half> %data) {
|
|
; GFX90A_GFX942-LABEL: name: global_atomic_fadd_v2f16_rtn
|
|
; GFX90A_GFX942: bb.0 (%ir-block.0):
|
|
; GFX90A_GFX942-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2
|
|
; GFX90A_GFX942-NEXT: {{ $}}
|
|
; GFX90A_GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX90A_GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; GFX90A_GFX942-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX90A_GFX942-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY1]], %subreg.sub1
|
|
; GFX90A_GFX942-NEXT: [[COPY3:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]]
|
|
; GFX90A_GFX942-NEXT: [[GLOBAL_ATOMIC_PK_ADD_F16_RTN:%[0-9]+]]:vgpr_32 = GLOBAL_ATOMIC_PK_ADD_F16_RTN killed [[COPY3]], [[COPY]], 0, 1, implicit $exec :: (load store syncscope("agent") seq_cst (s32) on %ir.ptr, addrspace 1)
|
|
; GFX90A_GFX942-NEXT: $vgpr0 = COPY [[GLOBAL_ATOMIC_PK_ADD_F16_RTN]]
|
|
; GFX90A_GFX942-NEXT: SI_RETURN_TO_EPILOG $vgpr0
|
|
%ret = atomicrmw fadd ptr addrspace(1) %ptr, <2 x half> %data syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
|
|
ret <2 x half> %ret
|
|
}
|
|
|
|
define amdgpu_ps <2 x half> @global_atomic_fadd_v2f16_saddr_rtn(ptr addrspace(1) inreg %ptr, <2 x half> %data) {
|
|
; GFX90A_GFX942-LABEL: name: global_atomic_fadd_v2f16_saddr_rtn
|
|
; GFX90A_GFX942: bb.0 (%ir-block.0):
|
|
; GFX90A_GFX942-NEXT: liveins: $sgpr0, $sgpr1, $vgpr0
|
|
; GFX90A_GFX942-NEXT: {{ $}}
|
|
; GFX90A_GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX90A_GFX942-NEXT: [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr1
|
|
; GFX90A_GFX942-NEXT: [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr0
|
|
; GFX90A_GFX942-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY1]], %subreg.sub1
|
|
; GFX90A_GFX942-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
|
|
; GFX90A_GFX942-NEXT: [[GLOBAL_ATOMIC_PK_ADD_F16_SADDR_RTN:%[0-9]+]]:vgpr_32 = GLOBAL_ATOMIC_PK_ADD_F16_SADDR_RTN killed [[V_MOV_B32_e32_]], [[COPY]], killed [[REG_SEQUENCE]], 0, 1, implicit $exec :: (load store syncscope("agent") seq_cst (s32) on %ir.ptr, addrspace 1)
|
|
; GFX90A_GFX942-NEXT: $vgpr0 = COPY [[GLOBAL_ATOMIC_PK_ADD_F16_SADDR_RTN]]
|
|
; GFX90A_GFX942-NEXT: SI_RETURN_TO_EPILOG $vgpr0
|
|
%ret = atomicrmw fadd ptr addrspace(1) %ptr, <2 x half> %data syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
|
|
ret <2 x half> %ret
|
|
}
|
|
|
|
define amdgpu_ps <2 x half> @global_atomic_fadd_v2f16_rtn_flat(ptr addrspace(1) %ptr, <2 x half> %data) {
|
|
; GFX90A_GFX942-LABEL: name: global_atomic_fadd_v2f16_rtn_flat
|
|
; GFX90A_GFX942: bb.0 (%ir-block.0):
|
|
; GFX90A_GFX942-NEXT: liveins: $vgpr0, $vgpr1, $vgpr2
|
|
; GFX90A_GFX942-NEXT: {{ $}}
|
|
; GFX90A_GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr2
|
|
; GFX90A_GFX942-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; GFX90A_GFX942-NEXT: [[COPY2:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX90A_GFX942-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sreg_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY1]], %subreg.sub1
|
|
; GFX90A_GFX942-NEXT: [[COPY3:%[0-9]+]]:vreg_64_align2 = COPY [[REG_SEQUENCE]]
|
|
; GFX90A_GFX942-NEXT: [[GLOBAL_ATOMIC_PK_ADD_F16_RTN:%[0-9]+]]:vgpr_32 = GLOBAL_ATOMIC_PK_ADD_F16_RTN killed [[COPY3]], [[COPY]], 0, 1, implicit $exec :: (load store syncscope("agent") seq_cst (s32) on %ir.ptr, addrspace 1)
|
|
; GFX90A_GFX942-NEXT: $vgpr0 = COPY [[GLOBAL_ATOMIC_PK_ADD_F16_RTN]]
|
|
; GFX90A_GFX942-NEXT: SI_RETURN_TO_EPILOG $vgpr0
|
|
%ret = atomicrmw fadd ptr addrspace(1) %ptr, <2 x half> %data syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
|
|
ret <2 x half> %ret
|
|
}
|
|
|
|
define amdgpu_ps <2 x half> @global_atomic_fadd_v2f16_saddr_rtn_flat(ptr addrspace(1) inreg %ptr, <2 x half> %data) {
|
|
; GFX90A_GFX942-LABEL: name: global_atomic_fadd_v2f16_saddr_rtn_flat
|
|
; GFX90A_GFX942: bb.0 (%ir-block.0):
|
|
; GFX90A_GFX942-NEXT: liveins: $sgpr0, $sgpr1, $vgpr0
|
|
; GFX90A_GFX942-NEXT: {{ $}}
|
|
; GFX90A_GFX942-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX90A_GFX942-NEXT: [[COPY1:%[0-9]+]]:sgpr_32 = COPY $sgpr1
|
|
; GFX90A_GFX942-NEXT: [[COPY2:%[0-9]+]]:sgpr_32 = COPY $sgpr0
|
|
; GFX90A_GFX942-NEXT: [[REG_SEQUENCE:%[0-9]+]]:sgpr_64 = REG_SEQUENCE [[COPY2]], %subreg.sub0, [[COPY1]], %subreg.sub1
|
|
; GFX90A_GFX942-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
|
|
; GFX90A_GFX942-NEXT: [[GLOBAL_ATOMIC_PK_ADD_F16_SADDR_RTN:%[0-9]+]]:vgpr_32 = GLOBAL_ATOMIC_PK_ADD_F16_SADDR_RTN killed [[V_MOV_B32_e32_]], [[COPY]], killed [[REG_SEQUENCE]], 0, 1, implicit $exec :: (load store syncscope("agent") seq_cst (s32) on %ir.ptr, addrspace 1)
|
|
; GFX90A_GFX942-NEXT: $vgpr0 = COPY [[GLOBAL_ATOMIC_PK_ADD_F16_SADDR_RTN]]
|
|
; GFX90A_GFX942-NEXT: SI_RETURN_TO_EPILOG $vgpr0
|
|
%ret = atomicrmw fadd ptr addrspace(1) %ptr, <2 x half> %data syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !0
|
|
ret <2 x half> %ret
|
|
}
|
|
|
|
!0 = !{}
|