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

[CI] Upgrade vllm to 0.9.1 (#1165)

Browse Source 
1. upgrade vllm to 0.9.1. 0.9.0 is not supported for main branch now.
keep doc to 0.9.0 until we release the first 0.9.1 release.
2. disable V0 test for PR
3. move actionlint check to lint job

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
This commit is contained in:
wangxiyuan
2025-06-11 16:33:11 +08:00
committed by GitHub
parent e46dc142bf
commit 4f5964420e
19 changed files with 72 additions and 320 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
vllm_ascend/patch/__init__.py
View File

@@ -24,9 +24,9 @@
# each worker's `__init__` function.
#
# Then in each kind of patch, there are three folders:
# - patch_0_9_0: contains the patches applied when vllm version is 0.9.0.
# - patch_0_9_1: contains the patches applied when vllm version is 0.9.1.
# - patch_main: contains the patches applied when vllm version is main branch.
# - patch_common: contains the patches applied in both 0.9.0 and main branch.
# - patch_common: contains the patches applied in both 0.9.1 and main branch.
#
# Once a new patch is added in vllm-ascend, please add the patch description into this file as well.
# ----------------------------------------------------------------------------------
@@ -35,17 +35,6 @@
# --------------------------------
# * Platform Patch:
# =================
# ** File: platform/patch_0_9_0/patch_distributed.py**
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# 1. `vllm.distributed.utils.stateless_init_torch_distributed_process_group()`
# Why:
# vllm distributed use gloo backend by default to initialize stateless process group, but we want to use hccl here
# How：
# Add hccl backend to the `stateless_init_torch_distributed_process_group`
# Related PR (if no, explain why):
# https://github.com/vllm-project/vllm/pull/18763
# Future Plan:
# Remove this patch once vllm is upgraded to 0.9.1
# ** File: platform/patch_common/patch_distributed.py**
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# 1. `vllm.distributed.parallel_state.destroy_model_parallel()`