Speed up PET backward indexing #828
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
frostedoyster
force-pushed
the
speed-up-pet-backward-indexing
branch
from
ead85ba to
6998ad3
Compare
frostedoyster
force-pushed
the
speed-up-pet-backward-indexing
branch
from
6998ad3 to
ff6ef85
Compare
Luthaf
reviewed
Oct 15, 2025
There was a problem hiding this comment.
I'll let @abmazitov comment on how much this makes sense for PET, but the code looks good to me
|
cscs-ci run |
|
This looks good to me, the branch is outdated so feel free to merge after updating to the latest changes in main |
abmazitov
approved these changes
Oct 17, 2025
|
cscs-ci run |
|
cscs-ci run |
|
cscs-ci run |
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Outdated suggestions cannot be applied.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.
As explained here pytorch/pytorch#41162, apparently having too many of the same index hurts performance dramatically in the backward of this operation (that we use for message passing). At the moment, our index for padding is always 0 and this creates a lot of duplicate indices.
Before:
After:
The speed improves (on my laptop's GPU) from 252 ms to 220 ms on a system of 5184 atoms.
📚 Documentation preview 📚: https://metatrain--828.org.readthedocs.build/en/828/