EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Refactor the ops PyTorch adapter，cleanup for csrc/torch_binding.cpp (#6732)

Browse Source 
### What this PR does / why we need it?
Refactor the ops PyTorch adapter，cleanup for csrc/torch_binding.cpp,
more details see
https://github.com/vllm-project/vllm-ascend/issues/6486

### Does this PR introduce _any_ user-facing change?
No

### How was this patch tested?
install the new package to test the new modification, here is the
result:


- vLLM version: v0.15.0
- vLLM main:
9562912cea

---------

Signed-off-by: liziyu <liziyu16@huawei.com>
Signed-off-by: wangxiaoteng <wangxiaoteng@huawei.com>
Signed-off-by: luomin2005 <luomin2005@huawei.com>
Co-authored-by: liziyu <56102866+liziyu179@users.noreply.github.com>
Co-authored-by: wangxiaoteng <wangxiaoteng@huawei.com>
This commit is contained in:
luomin2005
2026-02-24 09:12:43 +08:00
committed by GitHub
parent f0caeeadcb
commit f41eeeb11e
15 changed files with 1037 additions and 735 deletions
Download Patch File Download Diff File Expand all files Collapse all files

50
csrc/dispatch_layout/dispatch_layout_torch_adpt.h Normal file
View File

@@ -0,0 +1,50 @@
/*
* Copyright (c) Huawei Technologies Co., Ltd. 2026. All rights reserved.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef DISPATCH_LAYOUT_TORCH_ADPT_H
#define DISPATCH_LAYOUT_TORCH_ADPT_H
namespace vllm_ascend {
std::tuple<at::Tensor, at::Tensor, at::Tensor> get_dispatch_layout(const at::Tensor& topk_idx, int64_t num_experts,
int64_t num_ranks) {
TORCH_BIND_ASSERT(topk_idx.dim() == 2);
TORCH_BIND_ASSERT(topk_idx.is_contiguous());
TORCH_BIND_ASSERT(num_experts > 0);
const int num_tokens = topk_idx.size(0);
const int num_topk = topk_idx.size(1);
auto device = topk_idx.device();
auto num_tokens_per_expert = at::zeros({num_experts}, at::dtype(at::kInt).device(device));
auto num_tokens_per_rank = at::zeros({num_ranks}, at::dtype(at::kInt).device(device));
auto is_token_in_rank = at::zeros({num_tokens, num_ranks}, at::dtype(at::kInt).device(device));
EXEC_NPU_CMD(aclnnDispatchLayout,
topk_idx,
num_tokens,
num_ranks,
num_experts,
num_topk,
num_tokens_per_rank,
num_tokens_per_expert,
is_token_in_rank);
auto is_token_in_rank_bool = is_token_in_rank.to(at::kBool);
return std::make_tuple(num_tokens_per_rank, num_tokens_per_expert, is_token_in_rank_bool);
}
}
#endif