Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Refactoring of QR: try out pure stabilized Gram-Schmidt (split=1) and TS-QR (split=0) #1237

Open
mrfh92 opened this issue Oct 10, 2023 · 3 comments
Open

Refactoring of QR: try out pure stabilized Gram-Schmidt (split=1) and TS-QR (split=0) #1237

mrfh92 opened this issue Oct 10, 2023 · 3 comments
Assignees
@mrfh92
Labels
enhancement New feature or request linalg

Comments

@mrfh92
Copy link
Collaborator

mrfh92 commented Oct 10, 2023 •
edited

This issue is to suggest refactoring of the code for QR decomposition.
Currently, a quite elaborate algorithm is used for QR. The following suggestions might turn out to be actually faster or at least less prone to errors:

  • In the case of split=1, a straight-forward block-wise adaptation of stabilized Gram-Schmidt might be faster since this can exploit the highly optimized PyTorch-QR. This applies to split=1.

  • For split=0, a pure TS-QR might be an option.
@mrfh92 mrfh92 added the enhancement New feature or request label Oct 10, 2023
@mrfh92 mrfh92 self-assigned this Oct 10, 2023
@github-actions
Copy link
Contributor

Branch features/1237-Try_out_pure_stabilized_Gram-Schmidt_for_QR created!

@mrfh92 mrfh92 assigned mrfh92 and unassigned mrfh92 Oct 10, 2023
@github-actions GitHub Actions
Copy link
Contributor

This issue is stale because it has been open for 60 days with no activity.

@github-actions github-actions bot added the stale label Dec 18, 2023
@mrfh92 mrfh92 changed the title Try out pure stabilized Gram-Schmidt for QR Refactoring of QR: try out pure stabilized Gram-Schmidt (split=1) and TS-QR (split=0) Jan 22, 2024
This was referenced Jan 22, 2024
Closed
Closed
@mrfh92 mrfh92 added the linalg label Jan 22, 2024
@mrfh92
Copy link
Collaborator Author

mrfh92 commented Jan 22, 2024

For some of the issues with QR that lead to the need for extensive refactoring as proposed in this issue see, e.g., #1324 #1317

@mrfh92 mrfh92 removed the stale label Jan 26, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@mrfh92 mrfh92

Labels
enhancement New feature or request linalg
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@mrfh92