Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
41 changes: 39 additions & 2 deletions .github/workflows/aiter-release.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -67,6 +67,16 @@ on:
description: 'Use pytorch/manylinux2_28-builder ROCm image (AlmaLinux 8 + devtoolset, glibc 2.28). Produces wheels ABI-compatible with vLLM/Ubuntu 22 containers.'
type: boolean
default: false
torch_pin:
description: 'Optional torch version pin for the manylinux build (e.g. 2.10.0+rocm7.1). Empty = latest available for the detected ROCm flavor.'
type: string
required: false
default: ''
torch_index_url:
description: 'Optional override for the torch wheel index URL. Empty = auto-derive from the manylinux builder image tag (https://download.pytorch.org/whl/rocmX.Y).'
type: string
required: false
default: ''
workflow_call:
inputs:
release_type:
Expand Down Expand Up @@ -111,6 +121,16 @@ on:
type: boolean
required: false
default: false
torch_pin:
description: 'Optional torch version pin for the manylinux build (e.g. 2.10.0+rocm7.1). Empty = latest.'
type: string
required: false
default: ''
torch_index_url:
description: 'Optional torch index URL override. Empty = auto-derive from builder image tag.'
type: string
required: false
default: ''
outputs:
wheel_names:
description: 'Space-separated list of built wheel filenames'
Expand Down Expand Up @@ -145,6 +165,8 @@ jobs:
RELEASE_TYPE: ${{ inputs.release_type || github.event.inputs.release_type }}
ADD_DATE_STAMP: ${{ inputs.add_date_stamp || github.event.inputs.add_date_stamp }}
USE_MANYLINUX: ${{ inputs.use_manylinux || github.event.inputs.use_manylinux || (startsWith(matrix.docker_image, 'pytorch/manylinux') && 'true') || 'false' }}
TORCH_PIN: ${{ inputs.torch_pin || github.event.inputs.torch_pin }}
TORCH_INDEX_URL: ${{ inputs.torch_index_url || github.event.inputs.torch_index_url }}

steps:
- name: Checkout aiter repo
Expand Down Expand Up @@ -301,17 +323,32 @@ jobs:
IMG="${BUILD_DOCKER_IMAGE}"
ROCM_TAG="${IMG##*:}" # rocm7.2 / rocm7.1 / rocm7.0
ROCM_NUM="${ROCM_TAG#rocm}" # 7.2
TORCH_INDEX="https://download.pytorch.org/whl/rocm${ROCM_NUM}"
# Allow caller to override the torch wheel index (e.g. pin a release
# to a specific ROCm flavor's PyTorch ABI). Defaults preserve the
# legacy auto-derived behavior.
if [ -n "${TORCH_INDEX_URL}" ]; then
TORCH_INDEX="${TORCH_INDEX_URL}"
else
TORCH_INDEX="https://download.pytorch.org/whl/rocm${ROCM_NUM}"
fi
# Optional torch version pin (e.g. 2.10.0+rocm7.1). Empty = latest.
if [ -n "${TORCH_PIN}" ]; then
TORCH_SPEC="torch==${TORCH_PIN}"
else
TORCH_SPEC="torch"
fi
echo "Torch index: ${TORCH_INDEX}"
echo "Torch spec: ${TORCH_SPEC}"
docker exec \
-w /workspace \
-e PYBIN="${PYBIN}" \
-e TORCH_INDEX="${TORCH_INDEX}" \
-e TORCH_SPEC="${TORCH_SPEC}" \
aiter_build_${{ matrix.python_version }} \
bash -c '
set -e
${PYBIN}/pip install --upgrade --timeout=60 --retries=10 pip
${PYBIN}/pip install --timeout=60 --retries=10 --index-url "${TORCH_INDEX}" torch
${PYBIN}/pip install --timeout=60 --retries=10 --index-url "${TORCH_INDEX}" "${TORCH_SPEC}"
# flydsl publishes only manylinux_2_35 wheels which cannot install
# on AlmaLinux 8 (glibc 2.28). FlyDSL AOT pre-compilation in
# setup.py is wrapped in try/except and is skipped gracefully when
Expand Down
2 changes: 1 addition & 1 deletion 3rdparty/composable_kernel
Submodule composable_kernel updated 82 files
+87 −72 CMakeLists.txt
+16 −0 CMakePresets.json
+2 −0 Dockerfile.aiter
+2 −0 Dockerfile.fa
+2 −1 Dockerfile.pytorch
+3 −9 Jenkinsfile
+15 −0 README.md
+67 −22 dispatcher/python/ctypes_utils.py
+294 −0 dispatcher/tests/test_library_caching.py
+11 −4 example/40_conv2d_fwd_quantization/run_conv2d_fwd_perlayer_quantization_example.inc
+8 −2 example/ck_tile/01_fmha/CMakeLists.txt
+1 −1 example/ck_tile/01_fmha/README.md
+101 −12 example/ck_tile/01_fmha/codegen/ops/fmha_batch_prefill.py
+33 −2 example/ck_tile/01_fmha/fmha_fwd.hpp
+102 −10 example/ck_tile/01_fmha/fmha_fwd_runner.hpp
+0 −46 example/ck_tile/01_fmha/script/benchmark_fwd_v3.sh
+128 −0 example/ck_tile/52_cshuffle_lds/CMakeLists.txt
+61 −0 example/ck_tile/52_cshuffle_lds/README.md
+122 −0 example/ck_tile/52_cshuffle_lds/benchmark_cshuffle_lds.hpp
+100 −0 example/ck_tile/52_cshuffle_lds/benchmark_template.cpp.in
+3 −0 example/ck_tile/CMakeLists.txt
+5 −8 experimental/builder/include/ck_tile/builder/testing/conv/ck_tile.hpp
+6 −8 experimental/grouped_convolution_tile_instances/generate_instances.py
+2 −0 include/ck/host_utility/device_prop.hpp
+5 −4 include/ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_base.hpp
+6 −4 include/ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_selector.hpp
+10 −5 include/ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_v1.hpp
+1,216 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_data_multiple_d_xdl_cshuffle_v3.hpp
+2 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_dl.hpp
+1 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_explicit.hpp
+1 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_multiple_d_wmma_cshuffle_v3.hpp
+1 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_multiple_d_xdl_cshuffle.hpp
+1 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_two_stage_wmma_cshuffle_v3.hpp
+1 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_two_stage_xdl_cshuffle.hpp
+1 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_wmma_cshuffle_v3.hpp
+1 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_xdl_cshuffle.hpp
+17 −3 include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_xdl_cshuffle_v3.hpp
+167 −0 include/ck/tensor_operation/gpu/device/impl/device_grouped_gemm_fixed_nk_common.hpp
+7 −143 include/ck/tensor_operation/gpu/device/impl/device_grouped_gemm_multi_abd_wmma_fixed_nk.hpp
+14 −165 include/ck/tensor_operation/gpu/device/impl/device_grouped_gemm_multi_abd_xdl_fixed_nk.hpp
+6 −146 include/ck/tensor_operation/gpu/device/impl/device_grouped_gemm_wmma_fixed_nk.hpp
+13 −163 include/ck/tensor_operation/gpu/device/impl/device_grouped_gemm_xdl_fixed_nk.hpp
+11 −0 include/ck/tensor_operation/gpu/device/impl/split_k_utils.hpp
+3 −2 include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdl_cshuffle_common.hpp
+207 −121 include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdl_cshuffle_conv_v3.hpp
+38 −0 include/ck/tensor_operation/gpu/grid/gridwise_moe_gemm_blockscale.hpp
+4 −0 include/ck/tensor_operation/operator_transform/transform_conv_bwd_weight_to_gemm.hpp
+5 −0 include/ck/tensor_operation/operator_transform/transform_conv_bwd_weight_to_gemm_v2.hpp
+81 −0 include/ck_tile/core/arch/amd_buffer_addressing_builtins.hpp
+16 −7 include/ck_tile/core/tensor/tensor_descriptor.hpp
+168 −9 include/ck_tile/core/tensor/tile_scatter_gather.hpp
+14 −0 include/ck_tile/core/utility/type_traits.hpp
+1 −0 include/ck_tile/ops/fmha.hpp
+17 −0 include/ck_tile/ops/fmha/block/block_attention_kv_load_mode_enum.hpp
+10 −8 include/ck_tile/ops/fmha/kernel/fmha_batch_prefill_kernel.hpp
+248 −125 include/ck_tile/ops/fmha/pipeline/block_fmha_batch_prefill_pipeline_qr_ks_vs_async.hpp
+6 −0 include/ck_tile/ops/fmha/pipeline/block_fmha_pipeline_problem.hpp
+7 −2 include/ck_tile/ops/fmha/pipeline/tile_fmha_traits.hpp
+1 −1 include/ck_tile/ops/fused_moe/kernel/moe_sorting_kernel.hpp
+9 −4 include/ck_tile/ops/gemm/pipeline/gemm_pipeline_ag_bg_cr_comp_async_eight_waves_policy.hpp
+45 −0 include/ck_tile/utility/tile_load_store_microkernels.hpp
+85 −0 ...ibrary/tensor_operation_instance/gpu/grouped_conv_bwd_data/device_grouped_conv_bwd_data_xdl_v3_instance.hpp
+64 −4 ...ry/tensor_operation_instance/gpu/grouped_conv_bwd_weight/device_grouped_conv_bwd_weight_v3_xdl_instance.hpp
+4 −0 library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_backward_data.hpp
+28 −0 library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_backward_data_xdl.inc
+2 −0 library/src/tensor_operation_instance/gpu/grouped_conv2d_bwd_data/CMakeLists.txt
+49 −0 ...e/gpu/grouped_conv2d_bwd_data/xdl/device_grouped_conv2d_bwd_data_xdl_v3_nhwgc_gkyxc_nhwgk_bf16_instance.cpp
+49 −0 ...ce/gpu/grouped_conv2d_bwd_data/xdl/device_grouped_conv2d_bwd_data_xdl_v3_nhwgc_gkyxc_nhwgk_f16_instance.cpp
+24 −9 ...gc_gkzyxc_ndhwgk/device_grouped_conv3d_bwd_weight_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_default_pipev5_instance.cpp
+26 −9 ...dhwgc_gkzyxc_ndhwgk/device_grouped_conv3d_bwd_weight_xdl_ndhwgc_gkzyxc_ndhwgk_bf16_pad0_pipev5_instance.cpp
+25 −9 ...wgc_gkzyxc_ndhwgk/device_grouped_conv3d_bwd_weight_xdl_ndhwgc_gkzyxc_ndhwgk_f16_default_pipev5_instance.cpp
+25 −9 ...ndhwgc_gkzyxc_ndhwgk/device_grouped_conv3d_bwd_weight_xdl_ndhwgc_gkzyxc_ndhwgk_f16_pad0_pipev5_instance.cpp
+1 −0 python/ck4inductor/grouped_conv_fwd/op.py
+47 −1 script/cmake-ck-dev.sh
+28 −0 script/run_inductor_tests.sh
+7 −3 test/ck_tile/CMakeLists.txt
+6 −0 test/ck_tile/gemm_block_scale/CMakeLists.txt
+31 −0 test/ck_tile/gemm_block_scale/test_gemm_quant_abquant_eightwaves_padded_stride.cpp
+7 −2 test/ck_tile/gemm_block_scale/test_gemm_quant_fixtures.hpp
+11 −0 test/grouped_convnd_bwd_data/CMakeLists.txt
+258 −0 test/grouped_convnd_bwd_data/test_grouped_convnd_bwd_data_tile.cpp
+3 −3 tutorial/ck_tile/gemm/01_naive_gemm/README.md
Loading
Loading