Dan Hecht has posted comments on this change. ( http://gerrit.cloudera.org:8080/8363 )
Change subject: IMPALA-4456: Address scalability issues of qs_map_lock_ and client_request_state_map_lock_ ...................................................................... Patch Set 7: (4 comments) http://gerrit.cloudera.org:8080/#/c/8363/7/be/src/util/sharded-query-map-util.h File be/src/util/sharded-query-map-util.h: http://gerrit.cloudera.org:8080/#/c/8363/7/be/src/util/sharded-query-map-util.h@67 PS7, Line 67: struct MapShard { : std::unordered_map<TUniqueId, T> map_; : SpinLock map_lock_; : }; you should run a benchmark to verify, but you may want to explicitly align that to a cache line (by inheriting CacheLineAligned). http://gerrit.cloudera.org:8080/#/c/8363/7/be/src/util/sharded-query-map-util.h@94 PS7, Line 94: uint8_t see comment below about uint8_t http://gerrit.cloudera.org:8080/#/c/8363/7/be/src/util/sharded-query-map-util.h@122 PS7, Line 122: uint8_t why uint8_t? Since it doesn't live in memory, doesn't seem necessary to dictate the size. http://gerrit.cloudera.org:8080/#/c/8363/7/be/src/util/sharded-query-map-util.h@129 PS7, Line 129: std::unordered_map<TUniqueId, T>* map_; : SpinLock* map_lock_; This could be just a ShardedQueryMap::MapShard*, but up to you. -- To view, visit http://gerrit.cloudera.org:8080/8363 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I61089090e1095da45a8a64ed3ccc78bd310807f1 Gerrit-Change-Number: 8363 Gerrit-PatchSet: 7 Gerrit-Owner: Sailesh Mukil <sail...@cloudera.com> Gerrit-Reviewer: Dan Hecht <dhe...@cloudera.com> Gerrit-Reviewer: Philip Zeyliger <phi...@cloudera.com> Gerrit-Reviewer: Sailesh Mukil <sail...@cloudera.com> Gerrit-Comment-Date: Mon, 12 Feb 2018 22:41:53 +0000 Gerrit-HasComments: Yes
