Michael Ho has posted comments on this change. Change subject: IMPALA-4026: Implement double-buffering for BlockingQueue. ......................................................................
Patch Set 1: (2 comments) http://gerrit.cloudera.org:8080/#/c/4350/1//COMMIT_MSG Commit Message: Line 29: BlockingQueue to using POSIX pthread primitives instead of boost > This seems strange. I looked at ./boost_1_57_0/boost/thread/pthread/mutex.h In one of the backtrace from VTune, there was about 25% of the time of pthread_cond_wait() spent in the destructor of interruption_checker(). Similarly, there is an extra mutex_lock() in notify_one() and notify_all(). It's unclear what these extra checks in boost library are buying us. Line 45: | TARGETED-PERF(_300) | primitive_conjunct_ordering_1 | parquet / none / none | 10.72 | 9.84 | +8.95% | 2.57% | 0.53% | 1 | 3 | > Any idea why these regressed? Might be good to understand if there's someth Not sure if this is related to this patch. The nightly run from last night also shows similar (and in fact larger) level of regression: http://sandbox.jenkins.cloudera.com/job/impala-workload-runner-16node-cdh5/1763/ -- To view, visit http://gerrit.cloudera.org:8080/4350 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ib9f4cf351455efefb0f3bb791cf9bc82d1421d54 Gerrit-PatchSet: 1 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Michael Ho <k...@cloudera.com> Gerrit-Reviewer: Michael Ho <k...@cloudera.com> Gerrit-Reviewer: Mostafa Mokhtar <mmokh...@cloudera.com> Gerrit-Reviewer: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-HasComments: Yes