[Kernel] add custom op DispatchGmmCombineDecode (#4139)

#### What this PR does / why we need it?
add custom opapi DispatchGmmCombineDecode for A3, include kernel inpl,
python Api, pytest.

vLLM version: v0.11.0
vLLM main:
24d6314718


- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

Signed-off-by: wangqiankun <wangqiankun13@huawei.com>
Co-authored-by: wangqiankun <wangqiankun13@huawei.com>
Co-authored-by: wangxiyuan <wangxiyuan1007@gmail.com>
This commit is contained in:
GuoRen868
2025-12-06 17:33:14 +08:00
committed by GitHub
parent cb42564942
commit 4bd1030842
29 changed files with 7851 additions and 27 deletions

View File

@@ -163,6 +163,7 @@ cd ..
```
vllm-ascend will build custom operators by default. If you don't want to build it, set `COMPILE_CUSTOM_KERNELS=0` environment to disable it.
If you are building custom operators for Atlas A3, you should run `git submodule update --init --recursive` manually, or ensure your environment has Internet access.
:::
```{note}