[bugfix_v0.11.0-dev] layerwise D first plan (#3907)

Browse Source

### What this PR does / why we need it?
Refactored the layerwise code to send to the D node first, preventing
P-node hangs due to communication timeouts when DP > 1.
---------

Signed-off-by: nwpu-zxr <zhouxuerong2@huawei.com>
Signed-off-by: liziyu <liziyu16@huawei.com>
Signed-off-by: wangxiaoteng <wangxiaoteng@huawei.com>
Co-authored-by: nwpu-zxr <zhouxuerong2@huawei.com>
Co-authored-by: liziyu <liziyu16@huawei.com>

...

This commit is contained in:

wangxiaoteng888

2025-10-30 22:21:11 +08:00

committed by GitHub

parent d5a9aba03f

commit af7a56550b

5 changed files with 965 additions and 1356 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

859

vllm_ascend/distributed/mooncake_layerwise_connector.py

File diff suppressed because it is too large Load Diff