Make BaseChangeHermitian faster for large matrices #55
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Outdated suggestions cannot be applied.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.
... by performing row/column transformations directly instead of via matrix multiplication. This also reduces memory allocation and thus pressure on the garbage collector.
For comparison here are some timings using
TestBaseChangeHermitian
fromtst/basechange.tst
before and after this patch.Before:
After:
This is roughly on par with the orthogonal and symplectic cases:
Resolves #54 by @danielrademacher