Use builtin round instead of np.round for scalars. #14616
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
Compared to np.round, builtin round is 20x faster for python floats and 2x faster for numpy floats (it's admittedly unlikely that any of the places patched is an actual bottleneck). On Py2 builtin round and np.round were not equivalent because builtin round rounded halves away from zero, but now both round-to-even. Small misc. changes; the rearrangement in backend_agg.draw_text is to make the implementation more similar to draw_mathtext (just above).
tacaswell
approved these changes
Jun 25, 2019
|
Are there any cases where weird scalers (from pandas or one of the unit libraries) would be correctly handled by |
|
The name changes are because that example uses np.round so grepping found it... but it can't actually be changed to use builtin round there :) |
timhoffm
approved these changes
Jun 30, 2019
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Compared to np.round, builtin round is 20x faster for python floats and
2x faster for numpy floats (it's admittedly unlikely that any of the
places patched is an actual bottleneck).
On Py2 builtin round and np.round were not equivalent because builtin
round rounded halves away from zero, but now both round-to-even.
Small misc. changes; the rearrangement in backend_agg.draw_text is to
make the implementation more similar to draw_mathtext (just above).
PR Summary
PR Checklist