Get AXOL1TL model version from menu as of utm v0.12.0

[quinnanm]

PR description:

This PR changes the AXOL1TL emulator to fetch the model version from the menu xml rather than the configuration file (as is done currently here [1]). This allows for the model version included in the menu/firmware to be automatically rather than manually set in the emulator to correspond with the correct model version in cms-dist externals [2]. This was proposed to avoid model mismatch errors due to human error.

This PR requires the implementation of utm v0.12.0 [3], and should be merged once that is updated to be the default utm version/configuration. It is not compatible with previous utm versions due to changes in condition types (*if AXOL1TL conditions are included in the menu xml). This PR does not change the default AXOL1TL model version, which is currently v3 [4][5].

PR validation:

Changes were tested with the updated utm using the recipe described here [6]. A testing branch corresponding to this recipe is here [7]. This uses a test xml file for the new utm provided here [8], which is based on this example but with cicada and topo conditions removed [9]. Code-format changes were made, and it passes basic code-checks.

This update is scheduled to be presented in an upcoming Feb 26 L1DPG weekly meeting [10].

Many thanks to Bernhard Arnold and the l1t menu team for their work in adding model versioning capabilities to the utm, and to the L1DPG team and other contributors for their suggestions for this setup.

[1] https://github.com/cms-sw/cmssw/blob/master/L1Trigger/L1TGlobal/python/simGtStage2Digis_cfi.py#L19
[2] https://github.com/cms-hls4ml/AXOL1TL/tags
[3]https://globaltrigger.web.cern.ch/globaltrigger/release/utm/latest_doc/html/releaseNotes/v0_12_0.html
[4] https://github.com/cms-hls4ml/AXOL1TL/releases/tag/v3.0.0
[5] #43791
[6] https://codimd.web.cern.ch/s/q5TiuMfD7
[7]https://github.com/quinnanm/cmssw/tree/emulator_20feb24
[8]https://raw.githubusercontent.com/quinnanm/cmssw/emulator_20feb24/L1Trigger/L1TGlobal/data/Luminosity/startup/L1Menu_Collisions2024_AXO_TOPO_CICADA_v1_0_0_mod.xml
[9] https://cernbox.cern.ch/files/link/public/LeBbgndwFDjWoI5?tiles-size=1&items-per-page=100&view-mode=resource-table
[10] https://indico.cern.ch/category/2091/

[cmsbuild]

cms-bot internal usage

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-44054/38995

• This PR adds an extra 56KB to repository

[cmsbuild]

A new Pull Request was created by @quinnanm for master.

It involves the following packages:

• L1Trigger/L1TGlobal (l1)

@aloeliger, @epalencia, @cmsbuild can you please review it and eventually sign? Thanks.
@missirol, @Martin-Grunewald this is something you requested to watch as well.
@antoniovilela, @sextonkennedy, @rappoccio you are the release manager for this.

cms-bot commands are listed here

[Martin-Grunewald]

assign HLT

[Martin-Grunewald]

This affects HLT menus in ConfDb - L1TGlobalProducer - and will required a ConfDb parsing. So this PR (and its 14_0 backport) is time critical. In the meantime, to get this PR in, you need to code a customisation removing the AXOL1TL parameter here:
https://github.com/cms-sw/cmssw/blob/master/HLTrigger/Configuration/python/customizeHLTforCMSSW.py
(See the file for examples).
I assume the plan is to backport to 14_0?

[Martin-Grunewald]

assign hlt

[cmsbuild]

New categories assigned: hlt

@Martin-Grunewald,@mmusich you have been requested to review this Pull request/Issue and eventually sign? Thanks

[mmusich]

This PR requires the implementation of utm v0.12.0 [3], and should be merged once that is updated to be the default utm version/configuration.

For my edification, is that already the default utm version? how can one check that?

[epalencia]

This PR requires the implementation of utm v0.12.0 [3], and should be merged once that is updated to be the default utm version/configuration.

For my edification, is that already the default utm version? how can one check that?

The current version is 0.11.2: https://github.com/cms-sw/cmsdist/blob/IB/CMSSW_14_1_X/master/utm.spec

I just open a PR to update it to 0.12.0: cms-sw/cmsdist#9029

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-44054/39010

• This PR adds an extra 56KB to repository

[aloeliger]

L1 DQM uses fixed defined test versions of the L1 Menu

for my education, can you elaborate on this? what does "fixed" mean?

I'm thinking of the wrong thing. There are specific tests for defined unprescaled algorithms:

cmssw/DQM/L1TMonitor/python/L1TStage2uGTTiming_cfi.py

Lines 11 to 41 in 980274e

	unprescaledAlgoShortList = cms.untracked.vstring(
	"L1_SingleMu22_BMTF",
	"L1_SingleMu22_OMTF",
	"L1_SingleMu22_EMTF",
	"L1_SingleIsoEG28er1p5",
	"L1_SingleIsoEG32er2p5",
	"L1_SingleEG40er2p5",
	"L1_SingleEG60",
	"L1_SingleTau120er2p1",
	"L1_SingleJet180",
	"L1_ETMHF130",
	"L1_HTT360er",
	"L1_ETT2000"
	),
	prescaledAlgoShortList = cms.untracked.vstring(
	"L1_FirstCollisionInTrain",
	"L1_LastCollisionInTrain",
	"L1_IsolatedBunch",
	"L1_SingleMu0_BMTF",
	"L1_SingleMu0_OMTF",
	"L1_SingleMu0_EMTF",
	"L1_SingleEG10er2p5",
	"L1_SingleEG15er2p5",
	"L1_SingleEG26er2p5",
	"L1_SingleLooseIsoEG28er1p5",
	"L1_SingleJet35",
	"L1_SingleJet35er2p5",
	"L1_SingleJet35_FWD2p5",
	"L1_ETMHF100",
	"L1_HTT120er",
	"L1_ETT1600"

If I had to guess, my primary culprit would still be the menu change rather than anything this PR introduced (changing what bits mean what, causing emu mismatches). Let me talk to Caterina and see if I can understand what all went into this menu.

[artlbv]

Also, just to clarify, the DQM mismatches were not due to this PR itself as was essentially shown by the previous PR tests mentioned by Marco.

[aloeliger]

Okay, apparently this was a deeper L1T investigation stemming from the Repack. It's a known problem under investigation at this point.

[mmusich]

Okay, apparently this was a deeper L1T investigation stemming from the Repack. It's a known problem under investigation at this point.

would someone from the L1T team please open a gitHub issue such that there's trace of the investigations?
Thank you.

[cmsbuild]

REMINDER @antoniovilela, @sextonkennedy, @rappoccio: This PR was tested with #44389, please check if they should be merged together

[perrotta]

+alca

• Differences observed in L1TEMU DQM plots are discussed here above and do not pertain to AlCa
• This must be merged together with Update 2024 MC GTs with the L1 menu L1Menu_Collisions2024_v1_0_1_xml #44389 and Change hls4ml emulator extra version to avoid segfault cmsdist#9064

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @sextonkennedy, @antoniovilela, @rappoccio (and backports should be raised in the release meeting by the corresponding L2)
Notice This PR was tested with additional Pull Request(s), please also merge them if necessary: #44389

[rappoccio]

+1

[smuzaffar]

@antoniovilela, @rappoccio looks like this also needed cms-sw/cmsdist#9064 . As cmssw PR has been merged so I am merging cmsdist PR too

[antoniovilela]

@antoniovilela, @rappoccio looks like this also needed cms-sw/cmsdist#9064 . As cmssw PR has been merged so I am merging cmsdist PR too

Many thanks Shahzad.
Could you use the team mention (@cms-sw/orp-l2) which is easier to track down. I only saw this now.
Thanks.

[perrotta]

@smuzaffar @antoniovilela @rappoccio I remember a "revert" button that easily allowed to revert PRs... but perhaps it was only available for release managers, and I don't have those superpowers any more.

Following the outcomes of the joint operations meeting of 18th March, could any of you prepare the PRs to revert simultaneously:

• Get AXOL1TL model version from menu as of utm v0.12.0 #44054
• Update 2024 MC GTs with the L1 menu L1Menu_Collisions2024_v1_0_1_xml #44389

(While I don't think we need to revert also cms-sw/cmsdist#9064).

Thank you!

[mmusich]

Following the outcomes of the joint operations meeting of 18th March, could any of you prepare the PRs to revert simultaneously:

here it is: #44466

[mmusich]

@perrotta

I remember a "revert" button that easily allowed to revert PRs... but perhaps it was only available for release managers, and I don't have those superpowers any more.

for the record, the button is still there (at least I see it and I am not a release manager), but since there were conflicting commits on the same files I don't think the revert would have worked here in any case.

[perrotta]

for the record, the button is still there (at least I see it and I am not a release manager), but since there were conflicting commits on the same files I don't think the revert would have work here in any case.

Ah, ok: found it. I was looking for it in the wrong place
Thank you @mmusich