pepijnve commented on PR #23211: URL: https://github.com/apache/datafusion/pull/23211#issuecomment-4926485584
> I was thinking of some of the bitmap benchmarks we ran in arrow-rs where the state of the allocator (allocation order) mattered -- so that when run in isolation the tests didn't change but when run as the full suite the tests got slower Without an explicit allocator API that's going to be a tough one to control for. Ideally you would have an isolated arena per bench mark run so you can start from a clean slate every time. Would be useful to have that capability in arrow-rs and datafusion in general, but I think that's just a tad out of scope for this MR 😄 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
