apply sensitivity to counter numbers when explicitly specified

Before this commit, number of calls to counters were not applied any sensitivity. This stays true by default, but when a sensitivity is defined explicitly for a given counter, it will now be applied also to the number of calls

The reason of being of this MR is the be able to move more easily to DD4hep. As a matter of fact this move exercises the Kalman filter instabilities and despite identical geometries (but last digit differences in the double numbers), we get different results after the Kalman fit than with DD4Hep. Moreover, these differences also impact the number of calls to the counters of the Kalman fit, as the convergence can be different. Hence the need to be able to put a sensibility also on nb of counter calls.

added 1 commit

• e71574bf - apply sensitivity to counter numbers when explicitely specified

Compare with previous version

• [2020-06-22 14:44] Validation started with lhcb-dd4hep#562
• [2020-06-23 00:07] Validation started with lhcb-dd4hep#563
• [2020-06-23 10:30] Validation started with lhcb-dd4hep#564
• [2020-06-23 11:05] Validation started with lhcb-master-mr#974
• [2020-06-23 16:32] Validation started with lhcb-dd4hep#565
• [2020-06-24 00:14] Validation started with lhcb-dd4hep#566
• [2020-06-25 00:14] Validation started with lhcb-dd4hep#567
• [2020-06-26 00:09] Validation started with lhcb-gauss-dev#2430
• [2020-06-26 00:10] Validation started with lhcb-gaudi-head-py3#396
• [2020-06-26 00:11] Validation started with lhcb-lcg-dev3#1310
• [2020-06-26 00:11] Validation started with lhcb-gaussino#750
• [2020-06-26 00:11] Validation started with lhcb-lcg-dev4#1324
• [2020-06-26 00:12] Validation started with lhcb-dd4hep#568
• [2020-06-26 00:12] Validation started with lhcb-sanitizers#607
• [2020-06-26 00:17] Validation started with lhcb-lcg-test#75
• [2020-06-26 00:19] Validation started with lhcb-gaudi-head#2644
• [2020-06-26 00:20] Validation started with lhcb-head#2633
• [2020-06-26 01:00] Validation started with lhcb-sanitizers#607
• [2020-06-26 09:19] Validation started with lhcb-dd4hep#570
• [2020-06-26 09:30] Validation started with lhcb-dd4hep#571
• [2020-06-27 00:11] Validation started with lhcb-dd4hep#572
• [2020-06-27 00:21] Validation started with lhcb-gaudi-head#2645
• [2020-06-27 12:21] Validation started with lhcb-gaudi-head#2646
• [2020-06-28 00:09] Validation started with lhcb-gaudi-head#2647
• [2020-06-28 00:14] Validation started with lhcb-dd4hep#573
• [2020-06-29 00:08] Validation started with lhcb-gaudi-head#2648
• [2020-06-29 00:13] Validation started with lhcb-dd4hep#574

approved this merge request

assigned to @acasaisv

/ci-test --merge

assigned to @jonrob

resolved all threads

unassigned @acasaisv

can you clarify what you mean by "The reason of being of this MR is the move to DD4hep and the triggering of Kalman filter instabilities that goes with it" ? Do you mean the number of calls with DD4Hep is just different to with DetDesc, but still stable itself (i.e. always the same) or does the number actually fluctuate with DD4Hep. As if the latter we shouldn't be hiding that issue...

changed the description

changed title from apply sensitivity to counter numbers when explicitely specified to apply sensitivity to counter numbers when explicitly specified

added all-slots label

@sponce I am seeing a bunch of Moore/Brunel test failures due to counters diffs today. see e.g.

https://lhcb-nightlies.web.cern.ch/nightly/lhcb-gaudi-head/build/2644/

going to limit to one slot to test if this is the reason.

added lhcb-gaudi-head label and removed all-slots label

resolved all threads

removed lhcb-gaudi-head label

assigned to @sponce and unassigned @jonrob

added 1 commit

• 81a35e9d - apply sensitivity to counter numbers when explicitely specified

Compare with previous version