TrkExUtils - Escape from infinite loop

Hi everybody,

this MR provides a potential fix to move the stuck jobs in ATLASRECTS-6628 forward. The underlying reason is that the extrapolator tries to move a particle which just left ATLAS behind itself forward through ATLAS

RAWtoALL 09:04:48 MuonSegmentTagAlg.MuonSegmentT...AtlasExtrapolator     1     0   DEBUG create finalNextParameters  * TrackParameters on Surface
RAWtoALL 09:04:48  * loc1  : 9499.9955660
RAWtoALL 09:04:48  * loc2  : 0.1459753
RAWtoALL 09:04:48  * phi   : -2.3377817
RAWtoALL 09:04:48  * Theta : 2.6848200
RAWtoALL 09:04:48  * q/p   : 0.0000722
RAWtoALL 09:04:48  * charge: 1.0000000
RAWtoALL 09:04:48  * covariance matrix = 0xe412d930
RAWtoALL 09:04:48  * corresponding global parameters:
RAWtoALL 09:04:48  *    position  (x,  y,  z ) = (-6370.7449842, -7047.2352097, -22030.0000000)
RAWtoALL 09:04:48  *    momentum  (px, py, pz) = (-4238.3336794, -4397.3589019, -12427.6539981)
RAWtoALL 09:04:48 associated surface:

and an endless journey begins with not moving from the place. Given that changes in the extrapolator are always delicate, I tag @goblirsc, @christos, @sroe, @amorley, @gavrilen for another review, meaning that this MR deserves the label review-pending-expert. In the meantime, I'll launch a local test job over 10k events to ensure that nothing has changed there.

Tagging: @fmeloni, @oducu as reprocessing coordinators, @emoyse, @elmsheus, and @pberta as software coordinators

The philosopher of the day is kind of an outlier from the series but he nicely matches the theme of the day somehow.

added full-unit-tests label

This merge request affects 2 packages:

• Tracking/TrkExtrapolation/TrkExRungeKuttaPropagator
• Tracking/TrkExtrapolation/TrkExUtils

Affected files list will not be printed in this case

Adding @amorley ,@gavrilen as watchers

added 22.0-mc20 Tau Tracking review-pending-level-1 labels

added 1 deleted label and removed 22.0-mc20 Tau Tracking review-pending-level-1 labels

added 22.0-mc20 Tracking bugfix review-pending-expert labels

changed the description

CI Result SUCCESS (hash b07136e3)

	Athena
externals
cmake
make
required tests
optional tests

Full details available on this CI monitor view
Athena: number of compilation errors 0, warnings 0
For experts only: Jenkins output [CI-MERGE-REQUEST-CC7 40725]

Hi @goblirsc, @christos, @sroe, @amorley, @gavrilen, do you have any comments to this MR? I think it would be good to proceed with it.

Hi @pberta, I am by no means an extrapolator expert... from my attempts to reverse engineer what exactly the method is doing, this change seems OK at first glance, but I will try to take another look in the afternoon. If any of the real experts are available to comment, that would be highly appreciated! Also CCing @goetz @vcairo @gfacini

Two additional checks performed by @goblrisc and me from the muon & ID tracking sideAll_quick_plots.pdf, Valid_47445.pdf. Plots look good

Hi @christos, @sroe, @amorley, @gavrilen, do you have any comments to this MR? I am asking you just because Johannes tagged you at the beginning. It would be good to converge on this MR until the end of the week, so that it can be tested in the large-scale tests during the weekend.

Hi ,

I do not have any smart comment. I guess the issue is described and the solution is to be more careful on the check of being already there so as not to oscillate back and forth.

Hi @pberta,

the fact that the plots linked by @jojungge show a zero impact on the output is encouraging - maybe we can look into larger scale testing if this does not affect ongoing production?

Cheers, Max

added review-pending-level-1 urgent labels

removed review-pending-expert label

resolved all threads

Sounds like this has passed expert approval. From L1 side it also looks good. Approving.

Cheers, Volker (L1)

added review-approved label and removed review-pending-level-1 label

merged

mentioned in commit f2db74e8

added sweep:done label

picked the changes into the branch cherry-pick-f2db74e8a46-master with commit 6ab131af

Sweep summary
successful:

• master

(see https://aiatlas154.cern.ch/jenkins/job/auto_git_sweeps/1825/console for details)

mentioned in merge request !47690 (merged)

mentioned in commit 6ab131af

mentioned in commit a321c39e