llama : run all KQV ops on the CPU with no KV offload #5049
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
ggerganov
approved these changes
Jan 20, 2024
crasm
pushed a commit
that referenced
this pull request
Jan 23, 2024
jordankanter
pushed a commit
to jordankanter/llama.cpp
that referenced
this pull request
Feb 3, 2024
hodlen
pushed a commit
to hodlen/llama.cpp
that referenced
this pull request
Apr 1, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 change restores the behavior such that all the attention ops are run on the CPU when not offloading the KV. The net effect is that the amount of data that needs to be transferred between the CPU and GPU is much smaller, and performance should improve for large contexts, but for small contexts prompt/batch performance may be worse.
This should also fix nkvo with CUDA, but the underlying issue is still not known.