ramtorch: disable extensions by default for speedup on most systems #2548
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
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Outdated suggestions cannot be applied.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.
Ramtorch extensions for Embedding, RMSNorm, and other layers, for eg. Qwen Image, reduce the memory consumption from 25G to 12G, but on slower memory systems the speed drops by around 300% instead of 12%, from 14 seconds per step to 48 seconds per step.
On my 6000MHz DDR5 system, I don't observe such an extreme slowdown. But others have, and so, the option will be flipped by default so that the extensions are disabled.
This can be explicitly set to
falsein config, like the LTX-2 16G config example now does.