ATLASRECTS-5051 Part II
The cpu went down on ATLASRECTS-5051 , but there was a subtle drop in eff. This leave the Et cut at ~ 2.5 GeV which is was the main reason for the gain but loosens a tiny bit some of the other selection.
@amete and @rnaranjo I hope that after that we will an OK balance. I think we will need a bit more detailed tune to achieve the best eff with a good cpu timing which is part of https://its.cern.ch/jira/browse/ATLASRECTS-4772.
I think the issues are now well understood so we should try to have something even more reasonable soon