mridulm commented on code in PR #3487: URL: https://github.com/apache/celeborn/pull/3487#discussion_r2370842983
########## client/src/main/scala/org/apache/celeborn/client/commit/ReducePartitionCommitHandler.scala: ########## @@ -75,6 +75,7 @@ class ReducePartitionCommitHandler( private val stageEndShuffleSet = ConcurrentHashMap.newKeySet[Int]() private val inProcessStageEndShuffleSet = ConcurrentHashMap.newKeySet[Int]() private val shuffleMapperAttempts = JavaUtils.newConcurrentHashMap[Int, Array[Int]]() + private val shuffleToCompletedMappers = JavaUtils.newConcurrentHashMap[Int, Int]() Review Comment: Oh yes, completely agree ! Moving to native collections would not just improve memory utilization, but also have a good performance impact (boxing/unboxing/etc overheads). To be honest I had started on that a couple of times already - but each time abandoned due to time constraints. Moving to native collections would be a very good update to Celeborn ! -- 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]
