1de805ce0a
### What this PR does / why we need it?
During the prefill phase of Qwen3-Next and Qwen3.5, the
`torch.ops._C_ascend.causal_conv1d_fn` operator exhibits significant
performance bottlenecks. To address this, we have re-implemented the
optimization using `torch.ops._C_ascend.npu_causal_conv1d_custom`.
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
1 accuracy test
```
[2026-03-20 16:44:22,961] [ais_bench] [INFO] Start launch task state board ...
+-----------------------------+-----------+------------+-------------+----------+-------------------------------------------+---------------------+
| Task Name | Process | Progress | Time Cost | Status | Log Path | Extend Parameters |
+=============================+===========+============+=============+==========+===========================================+=====================+
| vllm-api-general-chat/gsm8k | 2918978 | NA | 0:00:01 | finish | logs/eval/vllm-api-general-chat/gsm8k.out | None |
+-----------------------------+-----------+------------+-------------+----------+-------------------------------------------+---------------------+
[2026-03-20 16:44:34,284] [ais_bench] [INFO] Evaluation tasks completed.
[2026-03-20 16:44:34,287] [ais_bench] [INFO] Summarizing evaluation results...
dataset version metric mode vllm-api-general-chat
--------- --------- -------- ------ -----------------------
gsm8k 271d0b accuracy gen 96.21
```
2 ut modify test
`pytest -sv
/home/c30006096/vllm-ascend/tests/e2e/nightly/single_node/ops/singlecard_ops/triton/test_causal_conv1d.py::test_ascend_causal_conv1d`
- vLLM version: v0.17.0
- vLLM main:
8b6325758c
Signed-off-by: wenba0 <3054239545@qq.com>
Signed-off-by: jiaojiao <56385650+wenba0@users.noreply.github.com>
61 lines
2.7 KiB
C++
61 lines
2.7 KiB
C++
/**
|
|
* Copyright (c) 2025 Huawei Technologies Co., Ltd.
|
|
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
|
|
* CANN Open Software License Agreement Version 2.0 (the "License").
|
|
* Please refer to the License for details. You may not use this file except in compliance with the License.
|
|
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
|
|
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
|
|
* See LICENSE in the root of the software repository for the full text of the License.
|
|
*/
|
|
|
|
/*!
|
|
* \file math_util.h
|
|
* \brief
|
|
*/
|
|
|
|
#ifndef TILING_MATMUL_MATH_UTIL_H
|
|
#define TILING_MATMUL_MATH_UTIL_H
|
|
|
|
#include <array>
|
|
#include <cstdint>
|
|
#include <vector>
|
|
#include <utility>
|
|
namespace matmul_tiling {
|
|
class MathUtil {
|
|
public:
|
|
static bool IsEqual(float leftValue, float rightValue);
|
|
template<typename T>
|
|
static auto CeilDivision(T num1, T num2) -> T
|
|
{
|
|
if (num2 == 0) {
|
|
return 0;
|
|
}
|
|
return static_cast<T>((static_cast<int64_t>(num1) + static_cast<int64_t>(num2) - 1) /
|
|
static_cast<int64_t>(num2));
|
|
}
|
|
template<typename T>
|
|
static auto Align(T num1, T num2) -> T
|
|
{
|
|
return CeilDivision(num1, num2) * num2;
|
|
}
|
|
static int32_t AlignDown(int32_t num1, int32_t num2);
|
|
static bool CheckMulOverflow(int32_t a, int32_t b, int32_t &c);
|
|
static int32_t MapShape(int32_t shape, bool roundUpFlag = true);
|
|
static void AddFactor(std::vector<int32_t> &dimsFactors, int32_t dim);
|
|
static void GetFactorCnt(const int32_t shape, int32_t &factorCnt, const int32_t factorStart,
|
|
const int32_t factorEnd);
|
|
static void GetFactorLayerCnt(const int32_t shape, int32_t &factorCnt, const int32_t factorStart,
|
|
const int32_t factorEnd);
|
|
static bool CheckFactorNumSatisfy(const int32_t dim);
|
|
static int32_t FindBestSingleCore(const int32_t oriShape, const int32_t mappedShape, const int32_t coreNum,
|
|
bool isKDim);
|
|
static void GetFactors(std::vector<int32_t> &factorList, int32_t srcNum, int32_t minFactor, int32_t maxFactor);
|
|
static void GetFactors(std::vector<int32_t> &factorList, int32_t srcNum, int32_t maxFactor);
|
|
static void GetBlockFactors(std::vector<int32_t> &factorList, const int32_t oriShape, const int32_t mpShape,
|
|
const int32_t coreNum, const int32_t maxNum);
|
|
static int32_t GetNonFactorMap(std::vector<int32_t> &factorList, int32_t srcNum, int32_t maxFactor);
|
|
static std::vector<std::pair<int, int>> GetFactorPairs(int32_t num);
|
|
static std::pair<int32_t, int32_t> DivideIntoMainAndTail(int32_t num, int32_t divisor);
|
|
};
|
|
} // namespace matmul_tiling
|
|
#endif // _MATH_UTIL_H_
|