zhengruifeng commented on PR #54555: URL: https://github.com/apache/spark/pull/54555#issuecomment-4003702991
> time_* benchmarks run with in-memory replay (io.BytesIO) to minimize file I/O noise in latency measurements. peakmem_* benchmarks run with temp-file replay to avoid keeping fully materialized input payloads in memory; temp files are explicitly removed in finally. @Yicong-Huang I think we should use the same way to generate input batches? can we reduce the memory footprint in this way: ``` batches = [batch1, batch2, batch3] iterator = random choose one batch from batches ``` -- 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]
