[1/N] Elastic EP Milestone 2 (#34861)

Signed-off-by: Yongji Wu <wuyongji317@gmail.com>
Signed-off-by: Itay Alroy <ialroy@nvidia.com>
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Signed-off-by: Ron Tourgeman <rtourgeman@nvidia.com>
Co-authored-by: Yongji Wu <wuyongji317@gmail.com>
Co-authored-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>
Co-authored-by: Ron Tourgeman <rtourgeman@nvidia.com>
This commit is contained in:
Itay Alroy
2026-02-28 04:46:42 +00:00
committed by GitHub
co-authored by Yongji Wu Tyler Michael Smith Ron Tourgeman
parent 90805ff464
commit dea268336f
53 changed files with 3612 additions and 1015 deletions
+6 -1
View File
@@ -7,6 +7,7 @@ import random
import torch
import torch.multiprocessing as mp
from vllm.config import VllmConfig, set_current_vllm_config
from vllm.distributed.parallel_state import (
init_distributed_environment,
)
@@ -42,7 +43,11 @@ def set_env_vars_and_device(env: dict[str, str]) -> None:
local_rank = os.environ["LOCAL_RANK"]
device = torch.device(f"cuda:{local_rank}")
torch.cuda.set_device(device)
init_distributed_environment()
# Create a minimal vllm config for init_distributed_environment
vllm_config = VllmConfig()
with set_current_vllm_config(vllm_config):
init_distributed_environment()
# Ensure each worker process has the same random seed
random.seed(42)