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
68 lines
3.7 KiB
MLIR
68 lines
3.7 KiB
MLIR
// RUN: mlir-opt %s -convert-gpu-to-rocdl='chipset=gfx942' | FileCheck %s --check-prefixes=CHECK,GFX9
|
|
// RUN: mlir-opt %s -convert-gpu-to-rocdl='chipset=gfx1201' | FileCheck %s --check-prefixes=CHECK,GFX12
|
|
|
|
gpu.module @test_module {
|
|
// CHECK-LABEL: func @subgroup_id()
|
|
func.func @subgroup_id() -> index {
|
|
// GFX12: %[[WAVEID:.+]] = rocdl.wave.id : i32
|
|
// GFX12: llvm.sext %[[WAVEID]] : i32 to i64
|
|
|
|
// GFX9-DAG: %[[IDX:.+]] = rocdl.workitem.id.x : i32
|
|
// GFX9-DAG: %[[IDY:.+]] = rocdl.workitem.id.y : i32
|
|
// GFX9-DAG: %[[IDZ:.+]] = rocdl.workitem.id.z : i32
|
|
// GFX9-DAG: %[[DIMX_I64:.+]] = llvm.call @__ockl_get_local_size(%[[C0:.+]]) : (i32) -> (i64 {llvm.range = #llvm.constant_range<i64, 1, 1025>})
|
|
// Yes, this is checking after the call that uses it. This prevents collisions with other 0s.
|
|
// GFX9-DAG: %[[C0]] = llvm.mlir.constant(0 : i32) : i32
|
|
// GFX9-DAG: %[[DIMX:.+]] = llvm.trunc %[[DIMX_I64]] overflow<nsw, nuw> : i64 to i32
|
|
// GFX9-DAG: %[[DIMY_I64:.+]] = llvm.call @__ockl_get_local_size(%[[C1:.+]]) : (i32) -> (i64 {llvm.range = #llvm.constant_range<i64, 1, 1025>})
|
|
// GFX9-DAG: %[[C1]] = llvm.mlir.constant(1 : i32) : i32
|
|
// GFX9-DAG: %[[DIMY:.+]] = llvm.trunc %[[DIMY_I64]] overflow<nsw, nuw> : i64 to i32
|
|
// GFX9: %[[Z_DY:.+]] = llvm.mul %[[DIMY]], %[[IDZ]] overflow<nsw, nuw>
|
|
// GFX9: %[[ZY:.+]] = llvm.add %[[IDY]], %[[Z_DY]] overflow<nsw, nuw>
|
|
// GFX9: %[[YZ_DX:.+]] = llvm.mul %[[DIMX]], %[[ZY]] overflow<nsw, nuw>
|
|
// GFX9: %[[ZYX:.+]] = llvm.add %[[IDX]], %[[YZ_DX]] overflow<nsw, nuw>
|
|
// GFX9: %[[WAVESZ:.+]] = rocdl.wavefrontsize : i32
|
|
// GFX9: %[[RES:.+]] = llvm.udiv %[[ZYX]], %[[WAVESZ]]
|
|
// GFX9: llvm.sext %[[RES]] : i32 to i64
|
|
%subgroupId = gpu.subgroup_id : index
|
|
func.return %subgroupId : index
|
|
}
|
|
|
|
// CHECK-LABEL: func @subgroup_id_with_upper_bound()
|
|
func.func @subgroup_id_with_upper_bound() -> index {
|
|
// GFX12: %[[WAVEID:.+]] = rocdl.wave.id range <i32, 0, 4> : i32
|
|
// GFX12: llvm.sext %[[WAVEID]] : i32 to i64
|
|
|
|
// Minimal check to ensure we don't set any bounds based on the subgroup ID bound
|
|
// since we don't know which thread ID they go on to.
|
|
// GFX9: rocdl.workitem.id.x : i32
|
|
// GFX9-DAG: llvm.call @__ockl_get_local_size({{.*}}) : (i32) -> (i64 {llvm.range = #llvm.constant_range<i64, 1, 1025>})
|
|
%subgroupId = gpu.subgroup_id upper_bound 4 : index
|
|
func.return %subgroupId : index
|
|
}
|
|
|
|
// CHECK-LABEL: func @subgroup_id_with_workgroup_sizes()
|
|
func.func @subgroup_id_with_workgroup_sizes() -> index
|
|
attributes {gpu.known_block_size = array<i32: 64, 4, 1>} {
|
|
// GFX12: %[[WAVEID:.+]] = rocdl.wave.id range <i32, 0, 4> : i32
|
|
// GFX12: llvm.sext %[[WAVEID]] : i32 to i64
|
|
|
|
// GFX9-DAG: %[[IDX:.+]] = rocdl.workitem.id.x range <i32, 0, 64> : i32
|
|
// GFX9-DAG: %[[IDY:.+]] = rocdl.workitem.id.y range <i32, 0, 4> : i32
|
|
// GFX9-DAG: %[[IDZ:.+]] = rocdl.workitem.id.z range <i32, 0, 1> : i32
|
|
// GFX9-DAG: %[[DIMX_I64:.+]] = llvm.mlir.constant(64 : i64) : i64
|
|
// GFX9-DAG: %[[DIMX:.+]] = llvm.trunc %[[DIMX_I64]] overflow<nsw, nuw> : i64 to i32
|
|
// GFX9-DAG: %[[DIMY_I64:.+]] = llvm.mlir.constant(4 : i64) : i64
|
|
// GFX9-DAG: %[[DIMY:.+]] = llvm.trunc %[[DIMY_I64]] overflow<nsw, nuw> : i64 to i32
|
|
// GFX9: %[[Z_DY:.+]] = llvm.mul %[[DIMY]], %[[IDZ]] overflow<nsw, nuw>
|
|
// GFX9: %[[ZY:.+]] = llvm.add %[[IDY]], %[[Z_DY]] overflow<nsw, nuw>
|
|
// GFX9: %[[YZ_DX:.+]] = llvm.mul %[[DIMX]], %[[ZY]] overflow<nsw, nuw>
|
|
// GFX9: %[[ZYX:.+]] = llvm.add %[[IDX]], %[[YZ_DX]] overflow<nsw, nuw>
|
|
// GFX9: %[[WAVESZ:.+]] = rocdl.wavefrontsize : i32
|
|
// GFX9: %[[RES:.+]] = llvm.udiv %[[ZYX]], %[[WAVESZ]]
|
|
// GFX9: llvm.sext %[[RES]] : i32 to i64
|
|
%subgroupId = gpu.subgroup_id upper_bound 4 : index
|
|
func.return %subgroupId : index
|
|
}
|
|
}
|