Skip to content

[Distributed] Use NVSHMEM symm-mem backend across nodes #48880

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
WoosukKwon wants to merge 1 commit into
base: main
Choose a base branch
from
+19 −0
Open

[Distributed] Use NVSHMEM symm-mem backend across nodes #48880

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions vllm/distributed/device_communicators/symm_mem.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,6 +87,25 @@ def __init__(
self.max_size = SYMM_MEM_ALL_REDUCE_MAX_SIZES[self.device_capability][
self.world_size
]
from vllm.distributed.parallel_state import in_the_same_node_as

if not all(in_the_same_node_as(self.group, source_rank=0)):
# Cross-node group (multi-node NVLink): the default CUDA backend
# exchanges CUDA IPC handles, which are same-node only. Select
# the NVSHMEM backend before the process's first symm-mem
# allocation so rendezvous maps fabric-reachable peer VAs.
try:
torch_symm_mem.set_backend("NVSHMEM")
logger.info_once(
"SymmMemCommunicator: multi-node group, using the "
"NVSHMEM symmetric-memory backend"
)
except RuntimeError as e:
logger.warning_once(
"SymmMemCommunicator: could not select the NVSHMEM "
"backend for a multi-node group: %s",
str(e),
)
try:
self.buffer = torch_symm_mem.empty(
self.max_size // self.dtype.itemsize,
Expand Down
Loading