[5321981] fix: Fix the Llama3.1 405B hanging issue. #5698
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
hyukn
force-pushed
the
fix/5321981
branch
2 times, most recently
from
6de8fa2 to
d694659
Compare
liji-nv
approved these changes
Jul 3, 2025
|
/bot run --disable-fail-fast --add-multi-gpu-test |
|
PR_Github #10770 [ run ] triggered by Bot |
litaotju
approved these changes
Jul 3, 2025
hyukn
changed the title
hyukn
changed the title
|
PR_Github #10770 [ run ] completed with state |
|
/bot run --disable-fail-fast --add-multi-gpu-test |
|
PR_Github #10827 [ run ] triggered by Bot |
|
PR_Github #10827 [ run ] completed with state |
Signed-off-by: Yukun He <23156053+hyukn@users.noreply.github.com>
|
/bot run |
|
PR_Github #10887 [ run ] triggered by Bot |
|
PR_Github #10887 [ run ] completed with state |
dc3671
pushed a commit
to dc3671/TensorRT-LLM
that referenced
this pull request
Jul 10, 2025
Correct the output shape of the fusedLayerNormPlugin. Signed-off-by: Yukun He <23156053+hyukn@users.noreply.github.com>
dc3671
pushed a commit
to dc3671/TensorRT-LLM
that referenced
this pull request
Jul 10, 2025
Correct the output shape of the fusedLayerNormPlugin. Signed-off-by: Yukun He <23156053+hyukn@users.noreply.github.com>
nvzhihanj
pushed a commit
to nvzhihanj/TensorRT-LLM
that referenced
this pull request
Jul 10, 2025
Correct the output shape of the fusedLayerNormPlugin. Signed-off-by: Yukun He <23156053+hyukn@users.noreply.github.com>
zhou-yuxin
pushed a commit
to zhou-yuxin/TensorRT-LLM
that referenced
this pull request
Jul 15, 2025
…NVIDIA#5698) (NVIDIA#5925) Signed-off-by: Yukun He <23156053+hyukn@users.noreply.github.com> Co-authored-by: Yukun He <23156053+hyukn@users.noreply.github.com> Signed-off-by: Yuxin <yuxinz@nvidia.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The output shapes of the
fusedLayerNormpluginfor nvFP4 are mismatched. This pollutes the barrier buffer for the one-shot allreduce kernel, which causes the hanging issue.This can also be the root cause of the accuracy issue as @zihaok recently mentioned. Because other data buffers are also messed up due to the out-of-range memory write.