Remove for loop over multiplicities in D_from_angles #517
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
Member
|
Hey thanks for the PR ! Can we have a comparison plot just to note down the speedup for future reference ? |
Contributor
Author
|
I tested the runtime of new and old version on a NVIDIA GeForce RTX 2080 Ti. I used the same random tensors of size 1000 for the angles while changing the irreps according to "{m}x0e+{m}x1o+{m}x2e" with m in [1, 2, 4, 8, 16, 32, 64]. 
|
Member
|
Thanks for the PR ! Should be available in the latest release 0.5.8 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
Remove the unnecessary for loop over multiplicities when using the function D_from_angles in the Irreps class.
Description
Line 743 in _irreps.py
direct_sum(*[ir.D_from_angles(alpha, beta, gamma, k) for mul, ir in self for _ in range(mul)])performs a for loop over multiplicities which in each iteration calculatesir.D_from_angleseven though the result is the same each time. The suggested change calculatesD_from_anglesonly once per irrep and repeats the resulting block according to the multiplicity.Motivation and Context
The change significantly increases the speed of the function especially in scenarios, where the number of multiplicities is high.
How Has This Been Tested?
The changes were tested by comparing the results of old and new implementation on an irreps object
Irreps("128x0e + 64x1o + 16x2e + 4x3o + 1x4e")for random tensors of angles and k and asserting that the resulting tensors are the same.Checklist: