-
Notifications
You must be signed in to change notification settings - Fork 4.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
set serial option for TH1::Fit in BeamSpotProducer #22227
Conversation
…uit thread local variables and IMT parallelization
The code-checks are being triggered in jenkins. |
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-22227/3376 |
A new Pull Request was created by @dan131riley (Dan Riley) for master. It involves the following packages: RecoVertex/BeamSpotProducer @perrotta, @ghellwig, @arunhep, @cerminar, @cmsbuild, @franzoni, @slava77, @lpernie can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
please test |
The tests are being triggered in jenkins. |
Comparison job queued. |
Comparison is ready Comparison Summary:
|
The code paths for the chi2 and likelihood evaluations are slightly different for the kSerial vs. kMultithread execution policies even single-threaded, but it's just sums and integrals, dunno to what extent there are random numbers involved. BinToBin is .996, with a fail threshold of 0.999999999999. I'm not really sure what to make of it, but I'd like to see this go in anyway to see if it fixes the crashes. |
+1
|
+1 |
This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @davidlange6, @slava77, @smuzaffar, @fabiocos (and backports should be raised in the release meeting by the corresponding L2) |
+1 |
Interactions between TMinuit thread local variables and implicit multi-threading in the evaluation of the fit are suspected to be a cause of crashes in AlcaBeamMonitor, details are in #21769 (long). This PR sets the SERIAL option to the calls to TH1::Fit, which turns off the use of IMT in the fit evaluation.