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

Remove VLLM_ASCEND_ENABLE_DENSE_OPTIMIZE (#5272)

Browse Source 
`VLLM_ASCEND_ENABLE_DENSE_OPTIMIZE` is only used together with
`VLLM_ASCEND_ENABLE_PREFETCH_MLP` which is useless totally. This PR
remove it.
- vLLM version: release/v0.13.0
- vLLM main:
ad32e3e19c

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
This commit is contained in:
wangxiyuan
2025-12-25 11:09:56 +08:00
committed by GitHub
parent 13cd6362c6
commit 2ae0bad96d
8 changed files with 7 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
vllm_ascend/ascend_forward_context.py
View File

@@ -102,8 +102,7 @@ def set_ascend_forward_context(
# TODO(rjg-lyh): refactor mlp weight prefetch method
# set for mlp weight prefetch
prefetch_mlp_enabled = envs_ascend.VLLM_ASCEND_ENABLE_DENSE_OPTIMIZE and \
envs_ascend.VLLM_ASCEND_ENABLE_PREFETCH_MLP and \
prefetch_mlp_enabled = envs_ascend.VLLM_ASCEND_ENABLE_PREFETCH_MLP and \
forward_context.layer_idx is not None and \
num_tokens is not None and num_tokens < 500
if prefetch_mlp_enabled: