xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-07-09 03:07:08 +08:00
Code Issues Packages Projects Releases Wiki Activity

updated

Browse Source 
Signed-off-by: Robert Shaw <robshaw@redhat.com>
This commit is contained in:
Robert Shaw 2025-07-07 01:13:20 +00:00
parent b835205d33
commit f16bf63877
1 changed files with 3 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
vllm/distributed/kv_transfer/kv_connector/v1/nixl_connector.py
View File

@ -329,17 +329,14 @@ class NixlConnectorWorker:
self.block_size = vllm_config.cache_config.block_size self.block_size = vllm_config.cache_config.block_size
# Agent. # Agent.
import os
num_workers = 32
# setting num_workers on the prefiller causes no notifs to be recved??? # setting num_workers on the prefiller causes no notifs to be recved???
# this is a hack to make sure we set num workers on the prefiller to 1. # this is a hack to make sure we set num workers on the prefiller to 1.
if os.getenv("VLLM_IS_PREFILL", "0") == "1": NUM_WORKERS = 32
num_workers = None logger.info(f"{NUM_WORKERS=}")
print(f"NUM_WORKERS: {num_workers=}")
self.nixl_wrapper = NixlWrapper(str(uuid.uuid4()), self.nixl_wrapper = NixlWrapper(str(uuid.uuid4()),
None, None,
num_workers=None, num_workers=None,
num_shared_workers=num_workers) num_shared_workers=NUM_WORKERS)
# Map of engine_id -> {rank0: agent_name0, rank1: agent_name1..}. # Map of engine_id -> {rank0: agent_name0, rank1: agent_name1..}.
self._remote_agents: dict[str, dict[int, str]] = defaultdict(dict) self._remote_agents: dict[str, dict[int, str]] = defaultdict(dict)