biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[XPU] check is_xccl_available before oneccl warmup (#39302)

Browse Source 
Signed-off-by: Xinyu Chen <xinyu1.chen@intel.com>
This commit is contained in:
Xinyu Chen
2026-04-10 03:42:17 +08:00
committed by GitHub
parent 2800706f06
commit a8134aef4e
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
vllm/v1/worker/xpu_worker.py
View File

@@ -86,7 +86,8 @@ class XPUWorker(Worker):
) )
# global all_reduce needed for overall oneccl warm up # global all_reduce needed for overall oneccl warm up
torch.distributed.all_reduce(torch.zeros(1).xpu()) if torch.distributed.is_xccl_available():
torch.distributed.all_reduce(torch.zeros(1).xpu())
# Set random seed. # Set random seed.
set_random_seed(self.model_config.seed) set_random_seed(self.model_config.seed)