gianm opened a new pull request, #19426: URL: https://github.com/apache/druid/pull/19426
In Dart, processing buffers are sliced up from the merge buffer. For stages that do not use all processing threads -- perhaps because they do not have enough inputs -- we can be more efficient with memory by slicing the merge buffer based on the actual number of processors, not the number of processing threads. This patch addresses it by deferring the choice of how many buffers are needed until the stage actually starts executing. At that point, it knows how many processors it will create. -- 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]
