You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It would be nice to have a standard suite of benchmarks -- particularly those things that are known performance bottlenecks -- so we can track their improvement or regression over time and provide data about the relative performance of different backends etc.
To increase visibility tagged as 'new-contributor-friendly' as this is a very self-contained project, but medium difficulty as it will involve sorting out how to set up ASV and how to find reliable hardware to run it on (our typical CI tricks won't work here as we need fixed resources underneath). Maybe pricing AWS for this?
We have some benchmarks here with results that end up here, from about 1.5.0 onwards. Those results are produced on my machine and I try to run them while I'm not using it, but there's not real guarantee of independence there.
We are working to get it running on the same dedicated hardware used for Pandas benchmarking.
Perhaps worth adding a benchmark on just the time to import matplotlib, or matplotlib.pyplot (but forcing the backend to agg, as that's going to be backend dependent).
This issue has been marked "inactive" because it has been 365 days since the last comment. If this issue is still present in recent Matplotlib releases, or the feature request is still wanted, please leave a comment and this label will be removed. If there are no updates in another 30 days, this issue will be automatically closed, but you are free to re-open or create a new issue if needed. We value issue reports, and this procedure is meant to help us resurface and prioritize issues that have not been addressed yet, not make them disappear. Thanks for your help!
Activity
tacaswell commentedon Aug 17, 2014
@mdboom seems like you already solved this https://github.com/spacetelescope/asv
Can we assign this to your mac-mini?
petehuang commentedon Dec 29, 2016
Hi,
This issue has been inactive for a while. Are there any updates? Any initial ideas that we have to help people start thinking about this? Thanks!
tacaswell commentedon Dec 29, 2016
To increase visibility tagged as 'new-contributor-friendly' as this is a very self-contained project, but medium difficulty as it will involve sorting out how to set up ASV and how to find reliable hardware to run it on (our typical CI tricks won't work here as we need fixed resources underneath). Maybe pricing AWS for this?
dopplershift commentedon Dec 29, 2016
Rackspace may also be an option; I think that's what powers https://speed.python.org/
jklymak commentedon Jan 25, 2018
Having the suite alone would be a useful first step. Automating it could come later.
QuLogic commentedon Jan 25, 2018
I am working on this; will try to publish something this weekend.
jklymak commentedon Jan 25, 2018
I'll bump the milestone though, as I don't think there is a higher rush. Just getting into master should be enough.
QuLogic commentedon Feb 4, 2018
We have some benchmarks here with results that end up here, from about 1.5.0 onwards. Those results are produced on my machine and I try to run them while I'm not using it, but there's not real guarantee of independence there.
We are working to get it running on the same dedicated hardware used for Pandas benchmarking.
anntzer commentedon Feb 18, 2018
Perhaps worth adding a benchmark on just the time to import matplotlib, or matplotlib.pyplot (but forcing the backend to agg, as that's going to be backend dependent).
tacaswell commentedon Feb 18, 2018
import / start up time has been something that @efiring is concerned about.
QuLogic commentedon Feb 20, 2018
There's a PR for that in asv: airspeed-velocity/asv#487 but it seems to have stalled.
QuLogic commentedon Oct 17, 2018
Waiting on asv-runner/asv-runner#2, though I think I need to update to asv 0.3 as well.
github-actions commentedon Mar 2, 2023
This issue has been marked "inactive" because it has been 365 days since the last comment. If this issue is still present in recent Matplotlib releases, or the feature request is still wanted, please leave a comment and this label will be removed. If there are no updates in another 30 days, this issue will be automatically closed, but you are free to re-open or create a new issue if needed. We value issue reports, and this procedure is meant to help us resurface and prioritize issues that have not been addressed yet, not make them disappear. Thanks for your help!