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

add pagedattention to support FULL_DECODE_ONLY. (#3102)

Browse Source 
### What this PR does / why we need it?
Calculate in advance the workspace memory size needed for the
PagedAttention operator to avoid deadlocks during resource cleanup. This
PR requires torch_npu version 0920 or newer.
### How was this patch tested?


- vLLM version: v0.11.0

---------

Signed-off-by: wangxiaoxin-sherie <wangxiaoxin7@huawei.com>
Co-authored-by: wangxiaoxin-sherie <wangxiaoxin7@huawei.com>
This commit is contained in:
XiaoxinWang
2025-10-10 08:50:33 +08:00
committed by GitHub
parent 1c2c72af8d
commit 579b7e5f21
5 changed files with 245 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
.github/workflows/_e2e_test.yaml vendored
View File

@@ -173,6 +173,7 @@ jobs:
if: ${{ inputs.type == 'full' }}
run: |
pytest -sv tests/e2e/multicard/test_data_parallel.py
pytest -sv tests/e2e/multicard/test_full_graph_mode.py
pytest -sv tests/e2e/multicard/test_expert_parallel.py
# external_launcher test is not stable enough. Fix it later
# pytest -sv tests/e2e/multicard/test_external_launcher.py