[CPU][Bugfix] Using custom allreduce for CPU backend (#15934)

Signed-off-by: jiang1.li <jiang1.li@intel.com>
This commit is contained in:
Li, Jiang
2025-04-02 07:46:47 -07:00
committed by GitHub
parent cefb9e5a28
commit 550b2801ad
8 changed files with 1013 additions and 16 deletions
+7
View File
@@ -1,5 +1,6 @@
# SPDX-License-Identifier: Apache-2.0
"""A CPU worker class."""
import os
from typing import Dict, List, Optional, Set, Tuple, Type
import torch
@@ -139,6 +140,8 @@ class CPUWorker(LocalOrDistributedWorkerBase):
self.local_rank = local_rank
self.rank = rank
vllm_config.parallel_config.rank = rank
self.distributed_init_method = distributed_init_method
self.is_driver_worker = is_driver_worker
@@ -217,6 +220,10 @@ class CPUWorker(LocalOrDistributedWorkerBase):
ret = torch.ops._C_utils.init_cpu_threads_env(self.local_omp_cpuid)
if ret:
logger.info(ret)
# Note: unique identifier for creating allreduce shared memory
os.environ["VLLM_DIST_IDENT"] = self.distributed_init_method.split(
":")[-1]
self.device = torch.device("cpu")
self.init_distributed_environment()
# Set random seed.