It would be useful also to join to Bugzilla thread dumps of 2.6 run (5 TD with 5 seconds intervall) . Call kill -3 <PID of JMETER>
Also add these options: -XX:+PrintGCApplicationConcurrentTime -XX:+PrintGCApplicationStoppedTime -XX:+PrintTenuringDistribution -XX:+PrintGCDetails -verbose:gc -XX:+PrintGCTimeStamps -XX:+PrintHeapAtGC -Xverbosegclog:<PATH TO A FILE NAME> And join the output file to Bugzilla. Thank you very much Philippe M. http://www.ubik-ingenierie.com On Wed, Feb 29, 2012 at 9:57 AM, ivan.rancati <ivan.ranc...@original1.net>wrote: > Hello, > good point about entering a Bugzilla issue. I would have to make a few > changes to the test plan and see how I can expose externally the server > under test, though, as there is confidential information. > > more details > - I am using GUI. The test plan has only one active listener (Aggregate > Report) > There are two CSV Data set configs for two files (each 99999 lines long) > The request is a plain http call, that uses values read from the files as > part of the GET request > - Non remote > - No 3rd party plugin > - I ran the test plan under release 2.6 and then under r1095243 again. top > showed a higher cpu usage with r1095243 > > thanks, > Ivan > > > > -- > View this message in context: > http://jmeter.512774.n5.nabble.com/Performance-difference-between-released-2-6-and-nighly-r1095243-tp5522066p5524406.html > Sent from the JMeter - User mailing list archive at Nabble.com. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org > For additional commands, e-mail: user-h...@jmeter.apache.org > > -- Cordialement. Philippe Mouawad.
