Community
Participate
Working Groups
You don't control what you can't *measure*. Performance is *key aspect* of model comparison. We used to work on performance but without constant checking we quickly loose the gain when changing code or integrating contributions. Let's work on having good measures of performances and scalability, and then we'll be able to check whether we have improvements or regressions. This framework (https://github.com/Obeo/fr.obeo.performance) has been initiated by Obeo guys to ease the measure and at some point, the reporting of these measures. It's EPL but it's not in Eclipse or has no CQ so far, meaning our performance test suite should be kept separate from the Eclipse repository for now.
I created a repo here :https://github.com/cbrun/emf.compare.performance
While this is certainly not perfect, https://hudson.eclipse.org/emfcompare/job/master-performance/ has been set up and keep track of performance with nice diagramming https://hudson.eclipse.org/emfcompare/job/master-performance/ws/performance/org.eclipse.emf.compare.tests.performance/index.html. Improvement requests to these tests should be done in a new bug.