Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
OpenDAS
vllm_cscc
Commits
d3ab240f
Unverified
Commit
d3ab240f
authored
Oct 28, 2025
by
Wentao Ye
Committed by
GitHub
Oct 28, 2025
Browse files
[Bug] Fix deepep low latency use nvlink by default (#27677)
Signed-off-by:
yewentao256
<
zhyanwentao@126.com
>
parent
94666612
Changes
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
2 additions
and
2 deletions
+2
-2
vllm/envs.py
vllm/envs.py
+2
-2
No files found.
vllm/envs.py
View file @
d3ab240f
...
@@ -205,7 +205,7 @@ if TYPE_CHECKING:
...
@@ -205,7 +205,7 @@ if TYPE_CHECKING:
VLLM_OBJECT_STORAGE_SHM_BUFFER_NAME
:
str
=
"VLLM_OBJECT_STORAGE_SHM_BUFFER"
VLLM_OBJECT_STORAGE_SHM_BUFFER_NAME
:
str
=
"VLLM_OBJECT_STORAGE_SHM_BUFFER"
VLLM_DEEPEP_BUFFER_SIZE_MB
:
int
=
1024
VLLM_DEEPEP_BUFFER_SIZE_MB
:
int
=
1024
VLLM_DEEPEP_HIGH_THROUGHPUT_FORCE_INTRA_NODE
:
bool
=
False
VLLM_DEEPEP_HIGH_THROUGHPUT_FORCE_INTRA_NODE
:
bool
=
False
VLLM_DEEPEP_LOW_LATENCY_ALLOW_NVLINK
:
bool
=
Fals
e
VLLM_DEEPEP_LOW_LATENCY_ALLOW_NVLINK
:
bool
=
Tru
e
VLLM_DEEPEP_LOW_LATENCY_USE_MNNVL
:
bool
=
False
VLLM_DEEPEP_LOW_LATENCY_USE_MNNVL
:
bool
=
False
VLLM_DBO_COMM_SMS
:
int
=
20
VLLM_DBO_COMM_SMS
:
int
=
20
GPT_OSS_SYSTEM_TOOL_MCP_LABELS
:
list
[
str
]
=
[]
GPT_OSS_SYSTEM_TOOL_MCP_LABELS
:
list
[
str
]
=
[]
...
@@ -1362,7 +1362,7 @@ environment_variables: dict[str, Callable[[], Any]] = {
...
@@ -1362,7 +1362,7 @@ environment_variables: dict[str, Callable[[], Any]] = {
# Allow DeepEP to use nvlink for internode_ll kernel, turn this on for
# Allow DeepEP to use nvlink for internode_ll kernel, turn this on for
# better latency on GB200 like system
# better latency on GB200 like system
"VLLM_DEEPEP_LOW_LATENCY_ALLOW_NVLINK"
:
lambda
:
bool
(
"VLLM_DEEPEP_LOW_LATENCY_ALLOW_NVLINK"
:
lambda
:
bool
(
int
(
os
.
getenv
(
"VLLM_DEEPEP_LOW_LATENCY_ALLOW_NVLINK"
,
"
0
"
))
int
(
os
.
getenv
(
"VLLM_DEEPEP_LOW_LATENCY_ALLOW_NVLINK"
,
"
1
"
))
),
),
# Allow DeepEP to use MNNVL (multi-node nvlink) for internode_ll kernel,
# Allow DeepEP to use MNNVL (multi-node nvlink) for internode_ll kernel,
# turn this for better latency on GB200 like system
# turn this for better latency on GB200 like system
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment