Add known dd4hep 'material corrections not supported' warning to exclusions

Adapts to Rec!3529 (merged) which adds a new known dd4hep warning.

added dd4hep testing labels

added ci-test-triggered label

• [2023-08-07 16:12] Validation started with lhcb-master-mr#8826
• [2023-08-07 16:32] Validation started with lhcb-master-mr#8827
• [2023-08-07 19:16] Validation started with lhcb-master-mr#8835
• [2023-08-08 10:19] Validation started with lhcb-master-mr#8836

added RTA label

changed title from Add known dd4hep 'material corrections no known' warning to exclusions to Add known dd4hep 'material corrections not supported' warning to exclusions

assigned to @jonrob

/ci-test Rec!3529 (merged)

added 5 commits

• 1d31fa90...7c0b54c8 - 4 commits from branch lhcb:master
• 189c1a5a - Add known dd4hep 'material corrections no known' warning to exclusions

Compare with previous version

/ci-test Rec!3529 (merged)

Started integration test build. Once done, check the results or the comparison to a reference build.

Throughput Test Moore_hlt2_fastest_reco: 483.2 Events/s -- change of -0.34% vs. reference

Throughput Test Moore_hlt2_pp_thor: 207.8 Events/s -- change of -0.53% vs. reference

Throughput Test Moore_spruce_all_lines: 199.6 Events/s -- change of 0.03% vs. reference . Total bandwidth 844.1 MB/s -- change of 0.03% vs. reference

Throughput Test Moore_hlt1_pp_default: 25920.5 Events/s -- change of 0.07% vs. reference

This "broke" some alignment tests: Alignment!414 (merged) . Some of the changes are non-trivial but I can't judge if they are expected (probably yes)

This MR didn't cause that. More Rec!3529 (merged)

Would it be better, instead of making the changes in Alignment!414 (merged), to instead extend the exclusions used there (so exactly what is done here) ?

I guess Alignment doesn't reuse the Moore exceptions, as if so they would have automatically received them via this MR.

Yes, you're right about the exclusions. One could also have them in Alignment but I was more worried about the real changes in the "tmf" test

yes, I indeed missed those diffs. My mistake.

I am quite confused by them though as the changes in Rec!3529 (merged) simply disable something which currently does not work with dd4hep. In Moore they where causing exceptions etc. I am confused how doing this could affect a task already running OK...

The alignment group needs to be asked to take a look I think

@lgrillo @mfontana Can you please ask the relevant tracking alignment expert(s) to look at the diffs in Alignment!414 (merged)

@rmatev How sure are you that Rec!3529 (merged) caused those diffs ? The new warnings, yes, but could something else have also gone in that caused them ?

The failing test is in build 8836 and 8835 (this set of MRs) while it's not in https://lhcb-nightlies.web.cern.ch/nightly/lhcb-master-mr/8833/ . It could be that something else got merged in between but I let you check that.

Well in that case I am willing to bet this set of MRs is not to blame, as I simply don't see how it could be possible. I'll let the alignment experts do the digging though.

only two (irrelevant) MRs in Moore were merged in between 8833 and 8835. You can check with

make for-each CMD='git log --reverse --merges --since="2023-08-07 17:00" --until="2023-08-07 19:30"'

So I'm pretty sure it's this MR.

What about other projects, LHCb or rec ?

Where I am coming from is the Rec MR addresses

Rec#500 (closed)

By disabling the use of the material locator with dd4hep, which if you use it results in

367 - TrackBestTrackCreator_66ad4814        FATAL  Standard std::exception is caught
400 - HLTControlFlowMgr                     FATAL Event failed in Node TrackBestTrackCreator/TrackBestTrackCreator_66ad4814 : NotImplemented at /workspace/build/Detector/InstallArea/x86_64_v2-centos7-gcc12-dbg/include/Core/Material.h:45

Maybe I am missing something but I just don’t see how this change, in a dd4hep build, could result in a change of output such as is seen.

Perhaps something in alignment is catching the exception and silently ignoring it. If so I would say the results its getting from the material locator is anyway a bit suspect…

resolved all threads

approved this merge request

resolved all threads

merged

mentioned in commit 5e82c414

mentioned in merge request Alignment!414 (merged)

added needs ref update label