Use invert event list to prefilter physics line when velo is closing

FYI @ascarabo @cagapopo

added RTA label

added 1 commit

• bbfa525e - add velo closed prefilter to all SMOG lines

Compare with previous version

resolved all threads

added rta-hotfix label

mentioned in merge request !1656 (closed)

mentioned in merge request !1658 (closed)

added 1 commit

• 30314152 - set DisableLinesDuringVPClosing to true by default

Compare with previous version

marked this merge request as ready

assigned to @msaur

added backport RTA/2024.06.03 label

added 35 commits

• 30314152...1dedfa99 - 32 commits from branch 2024-patches
• 41a794f2 - Use invert event list to prefilter physics line when velo is closing
• 6aa4a49e - add velo closed prefilter to all SMOG lines
• db397789 - set DisableLinesDuringVPClosing to true by default

Compare with previous version

/ci-test

added ci-test-triggered label

• [2024-06-14 19:52] Validation started with lhcb-2024-patches-mr#769
• [2024-06-15 13:37] Validation started with lhcb-2024-patches-mr#774
• [2024-06-16 12:50] Validation started with lhcb-2024-patches-mr#779
• [2024-06-16 14:58] Automatic merge failed in [lhcb-2024-patches-mr#780](https://lhcb-nightlies.web.cern.ch/nightly/lhcb-2024-patches-mr/780/Allen/checkout

added hlt1-throughput-decreased label

there is a small throughput decrease, namely 2.5% for A500.

there are also physics differences for the sequence forward_then_matching_and_downstream, eg

+++ efficiency_Beam6800GeV-expected-2024-MagDown-nu7.6_Bs2PhiPhiMD_nSciFiHits_hlt1_pp_forward_then_matching_and_downstream_validation_a5000.txt	2024-06-15 01:14:28.535918900 +0200
@@ -377,12 +377,12 @@
 Hlt1SMOG2SingleMuon:                                0           0
 Hlt1SMOG2L0Toppi:                                   0           0
 
-Total decisions:      581
-Total tracks:         1204
+Total decisions:      584
+Total tracks:         1208
 Total calos clusters: 20
-Total SVs:            1117
-Total hits:           31086
-Total stdinfo:        14761
+Total SVs:            1126
+Total hits:           31194
+Total stdinfo:        14832

and similar for the same sequence in other GPUs. The other sequences look unaffected.

I was expecting no impact at all on the physics from this MR. Any idea @ahennequ @cagapopo ? Are the references up to date @msaur ?

There are some throughput decreases as well:

sequence hlt1_pp_rich_no_ut over dataset Run_0000248711_HLT20840_20221011-113809-426: % change: -4.19% 
sequence hlt1_pp_no_ut over dataset Beam6800GeV-expected-2024-MagDown-nu7.6_MinBiasMD: % change: -3.12%

is that expected or assumed to be just a fluctuation?

/ci-test

mentioned in issue #558

/ci-test !1644 (merged)

removed hlt1-throughput-decreased label

mentioned in merge request !1680 (merged)

mentioned in issue MooreOnline#79 (closed)

mentioned in commit c0142a5c

mentioned in commit 2e8e21c7

mentioned in merge request !1681 (merged)

mentioned in commit Moore@d6477fef

mentioned in merge request Moore!3571 (merged)

mentioned in commit Moore@573e221b

Reference updated: !1681 (merged) ; Moore!3571 (merged)

added 1 commit

• e566c292 - Update CI references

Compare with previous version

resolved all threads

mentioned in commit da417429

merged

mentioned in commit 2040993e

mentioned in commit 566a8ffa

mentioned in commit Moore@9fe0c4bf

mentioned in commit Moore@127d33dd

mentioned in commit 00e1638d

mentioned in merge request !1668 (merged)

mentioned in merge request Moore!3554 (merged)