Files
xc-llm-ascend/csrc/build_aclnn.sh
Chen Chen ad0607f900 add dispatch_gmm_combine kernel (#3532)
### What this PR does / why we need it?

This PR introduces the Ascend implementation of the
`dispatch_ffn_combine` kernel and wires it into the vLLM-Ascend runtime,
together with follow‑up fixes to ensure the kernel builds and runs
correctly in CI.

- Add full host and device implementation of the `dispatch_ffn_combine`
kernel under `csrc/dispatch_ffn_combine`, including tiling logic, MOE
routing helpers, and kernel utilities for quantized FFN dispatch.
- Integrate the new kernel with the PyTorch binding
(csrc/torch_binding.cpp, csrc/torch_binding_meta.cpp) and the Ascend
runtime (vllm_ascend/ascend_forward_context.py,
vllm_ascend/worker/model_runner_v1.py).
- Extend fused MoE communication and token dispatch support in
`vllm_ascend/ops/fused_moe`, adding methods/utilities needed by the new
dispatch path.
- Update quantization logic in vllm_ascend/quantization/w8a8_dynamic.py
to support the new FFN dispatch flow.
- Fix kernel build issues by adjusting `csrc/build_aclnn.sh`, CMake
configuration, and include/namespace usage in the new kernel files.
- Add an end‑to‑end nightly test
`tests/e2e/nightly/ops/test_dispatch_ffn_combine.py` and helper
utilities in `vllm_ascend/utils.py` to validate the new kernel.

### Does this PR introduce _any_ user-facing change?

### How was this patch tested?


- vLLM version: v0.12.0
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.12.0

---------

Signed-off-by: mojave2 <chenchen145@huawei.com>
Co-authored-by: wangxiyuan <wangxiyuan1007@gmail.com>
2025-12-04 23:00:59 +08:00

60 lines
1.9 KiB
Bash

#!/bin/bash
ROOT_DIR=$1
SOC_VERSION=$2
git config --global --add safe.directory "$ROOT_DIR"
if [[ "$SOC_VERSION" =~ ^ascend310 ]]; then
# ASCEND310P series
# currently, no custom aclnn ops for ASCEND310 series
# CUSTOM_OPS=""
# SOC_ARG="ascend310p"
exit 0
elif [[ "$SOC_VERSION" =~ ^ascend910b ]]; then
# ASCEND910B (A2) series
CUSTOM_OPS="grouped_matmul_swiglu_quant_weight_nz_tensor_list;lightning_indexer;sparse_flash_attention;dispatch_ffn_combine"
SOC_ARG="ascend910b"
elif [[ "$SOC_VERSION" =~ ^ascend910_93 ]]; then
# ASCEND910C (A3) series
CUSTOM_OPS="grouped_matmul_swiglu_quant_weight_nz_tensor_list;lightning_indexer;sparse_flash_attention;dispatch_ffn_combine"
SOC_ARG="ascend910_93"
else
# others
# currently, no custom aclnn ops for other series
exit 0
fi
git submodule init
git submodule update
# For the compatibility of CANN8.5 and CANN8.3: copy and modify moe_distribute_base.h
file_path=$(find /usr/local/Ascend/ascend-toolkit -name "moe_distribute_base.h" 2>/dev/null | head -n1)
if [ -z "$file_path" ]; then
echo "cannot find moe_distribute_base.h file in CANN env"
exit 1
fi
SCRIPT_DIR=$(cd "$(dirname "$0")" && pwd)
TARGET_DIR="$SCRIPT_DIR/dispatch_ffn_combine/op_kernel/utils/"
TARGET_FILE="$TARGET_DIR/$(basename "$file_path")"
echo "*************************************"
echo $file_path
echo "$TARGET_DIR"
cp "$file_path" "$TARGET_DIR"
sed -i 's/struct HcclOpResParam {/struct HcclOpResParamCustom {/g' "$TARGET_FILE"
sed -i 's/struct HcclRankRelationResV2 {/struct HcclRankRelationResV2Custom {/g' "$TARGET_FILE"
# build custom ops
cd csrc
rm -rf build output
echo "building custom ops $CUSTOM_OPS for $SOC_VERSION"
bash build.sh -n $CUSTOM_OPS -c $SOC_ARG
# install custom ops to vllm_ascend/_cann_ops_custom
./output/CANN-custom_ops*.run --install-path=$ROOT_DIR/vllm_ascend/_cann_ops_custom