Fix meta backend tensor reads for split tensors during state serialization#22063
Merged
JohannesGaessler merged 1 commit intoApr 18, 2026
Merged
Conversation
github-actions
Bot
added
the
ggml
JohannesGaessler
approved these changes
Apr 17, 2026
pwilkin
approved these changes
Apr 18, 2026
Member
|
Can confirm it now works. Performance is slightly worse at 93 t/s tg vs 110 on |
Member
|
Actually, never mind, was missing NCCL. With NCCL installed, performance is up at 123 t/s. |
samuraieng
pushed a commit
to samuraieng/llama.cpp
that referenced
this pull request
Apr 19, 2026
mengqin
pushed a commit
to mengqin/llama.cpp
that referenced
this pull request
Apr 20, 2026
ArberSephirotheca
pushed a commit
to ArberSephirotheca/llama.cpp
that referenced
this pull request
Apr 21, 2026
arthw
pushed a commit
to arthw/llama.cpp
that referenced
this pull request
Apr 23, 2026
rsenthilkumar6
pushed a commit
to rsenthilkumar6/llama.cpp
that referenced
this pull request
May 1, 2026
jimbothigpen
pushed a commit
to jimbothigpen/frankenturbo2
that referenced
this pull request
May 2, 2026
ljubomirj
pushed a commit
to ljubomirj/llama.cpp
that referenced
this pull request
May 6, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR fixes a crash when saving recurrent state with tensor-split models using the meta backend. The previous code assumed that a tensor read would always map to a single segment, which is not always true when -sm tensor is enabled. The fix handles multi-segment tensor reads correctly instead of hitting the split_state.n_segments == 1 assertion. This should allow checkpoint/state serialization to work reliably with tensor-parallel CUDA setups. Fixes #22058