Skip to content

[sglang-miles] Add SGLANG_DISABLE_MULTIMEM_AG to force the NCCL all-gather path #36110

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

Merged
yueming-yuan merged 1 commit into from
Aug 24, 2026
Merged

[sglang-miles] Add SGLANG_DISABLE_MULTIMEM_AG to force the NCCL all-gather path #36110

Show file tree
Hide file tree
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
4 changes: 4 additions & 0 deletions python/sglang/srt/distributed/device_communicators/triton_symm_mem_ag.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,8 @@
import triton
import triton.language as tl

from sglang.srt.environ import envs

logger = logging.getLogger(__name__)

# Each thread moves _NUMEL_PER_THREAD bf16 via one 128-bit multimem op; the
Expand Down Expand Up @@ -460,6 +462,8 @@ def __init__(
):
self._max_tokens = int(max_tokens)
self._skip_entry_sync = skip_entry_sync
if envs.SGLANG_DISABLE_MULTIMEM_AG.get():
enabled = False
# None => always NCCL; _UNINIT => build on first eager call.
self._state = self._UNINIT if enabled else None
if self._state is self._UNINIT:
Expand Down
1 change: 1 addition & 0 deletions python/sglang/srt/environ.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1107,6 +1107,7 @@ class Envs:
# Symmetric Memory
SGLANG_SYMM_MEM_PREALLOC_GB_SIZE = EnvInt(-1)
SGLANG_DEBUG_SYMM_MEM = EnvBool(False)
SGLANG_DISABLE_MULTIMEM_AG = EnvBool(False)

# Aiter
SGLANG_USE_AITER_FP8_PER_TOKEN = EnvBool(False)
Expand Down
Loading