Skip to content

Conversation

@daharoni
Copy link
Member

@daharoni daharoni commented Dec 15, 2025

Follows #38

Summary

Without making any algorithm changes, find places to speed up sparse matrix code

Changes

  • solver.py
    • Use row-scaling: M.T @ W^2 @ M == (W @ M).T @ (W @ M). So we now do Mw = M.multiply(w[:, None]) and then P = Mw.T @ Mw. ~12% speed up on short profiling test.

@daharoni daharoni linked an issue Dec 15, 2025 that may be closed by this pull request
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

scipy/sparse/_compressed.py is where a lot of compute time goes

2 participants