E523 BUG/REF: Use lru_cache instead of NUMBA_FUNC_CACHE by mroeschke · Pull Request #46086 · pandas-dev/pandas · GitHub
[go: up one dir, main page]

Skip to content

Conversation

@mroeschke
Copy link
Member

@mroeschke mroeschke added Bug numba numba-accelerated operations labels Feb 20, 2022
@mroeschke mroeschke added this to the 1.5 milestone Feb 21, 2022
Copy link
Contributor
@jreback jreback left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

assume this has neutral perf impact

starts, ends, sorted_index, sorted_data = self._numba_prep(df)
aggregator = executor.generate_shared_aggregator(
func, engine_kwargs, numba_cache_key_str
func, *get_jit_arguments(engine_kwargs)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can we change get_jit_argument to return a dict and splat here? e.g. always keyword args

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure. Done.

@mroeschke
Copy link
Member Author

assume this has neutral perf impact

Yup, additionally this actually reduces some unnecessary jitting when e.g. going from rolling.mean to expanding.mean so that might be a slight performance bump in those odd cases

@jreback jreback merged commit 696a8e9 into pandas-dev:main Feb 27, 2022
@mroeschke mroeschke deleted the bug/numba_caching branch February 28, 2022 00:12
yehoshuadimarsky pushed a commit to yehoshuadimarsky/pandas that referenced this pull request Jul 13, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Bug numba numba-accelerated operations

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants

0