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
TLS cleanup #4293
TLS cleanup #4293
Conversation
A new Pull Request was created by @VinInn (Vincenzo Innocente) for CMSSW_7_2_X. TLS cleanup It involves the following packages: CondFormats/EcalObjects @apfeiffer1, @nclopezo, @cmsbuild, @diguida, @rcastello, @StoyanStoynev, @slava77, @ggovi, @Degano can you please review it and eventually sign? Thanks. |
Pull request #4293 was updated. @apfeiffer1, @nclopezo, @cmsbuild, @diguida, @rcastello, @StoyanStoynev, @slava77, @ggovi, @Degano can you please check and sign again. |
The following comparisons:https://cmssdt.cern.ch/SDT/jenkins-artifacts/baseLineComparisons/CMSSW_7_2_X_2014-06-17-0200+4293/2881/25.0_TTbar+TTbar+DIGI+RECO+HARVEST+ALCATT/comparisonResults/Tracking_TrackParameters_highPurityTracks_pt_1_GeneralProperties.htmlare failing. Is this expected? |
25.0 from jenkins includes resim. this is occasionally not reproducible, the random sequence diverges at some point. |
@StoyanStoynev For what EcalCond*Container is concerned: they is no chance that anything else than a EBAR or ECAL is passed there... @argiro to comment further |
All tests I made were fine with one exception - wf 19 (SingleGammaPt35). I see a difference in conversion related variables. It originates from this as far as I see: < recoConversions_allConversions__RECO. 4902 1128.8
|
What do you mean when you say "I used 33d013d"? Can you try in last night IB, with and without the PR? |
This looks like one of the cases I've seen in I think now that this code/test should be run with valgrind. On 6/19/14, 6:04 AM, StoyanStoynev wrote:
Vyacheslav (Slava) Krutelyov |
@slava77 are we sure the diff comes from RECO and not from DIGI? |
previously the variable is somewhere unrelated to this PR, no? if so, we need to work in parallel. On Jun 19, 2014, at 3:17 PM, Slava Krutelyov notifications@github.com
|
The PR deals with trajectory filtering and cleaning (by shared hits) code - I would suppose it is related to the conversion reconstruction. |
@VinInn , I mentioned there was small change in the event size - the branch recoConversions_allConversions__RECO was affected (this is step3) and there were no difference at all at step2 (digis). It should be RECO |
Thanks for the investigation. |
+1 |
Bypassing alca and DB. Complain if not ok. |
For the record, when running under valgrind
can someone look into it? |
most probably irrelevant On 19 Jun, 2014, at 4:45 PM, Giulio Eulisse notifications@github.com wrote:
|
I opened an issue in my area, feel free to continue to comment there on the allConversions reproducibility |
On 19 Jun 2014, at 11:27, Vincenzo Innocente notifications@github.com wrote:
sorry I was traveling . I concur with Vincenzo. It’s very unlikely that the 1 conversion difference is from this piece of code
|
Backport TLS fixes from #4293 to CMSSW_7_1_X
removed a huge number of useless TLS from ECAL CondContainer(s)
reduced by a factor 2 in TrajectoryCleaner and TrajectoryFilter