-
Notifications
You must be signed in to change notification settings - Fork 4.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add L1 Upgrade Workflow: 0.78 #43271
Conversation
Should run for Digi or DigiTrigger steps, and on 2026 keys. Will include the Phase 2 GT emulator
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-43271/37674
|
A new Pull Request was created by @aloeliger (Andrew Loeliger) for master. It involves the following packages:
@miquork, @AdrianoDee, @srimanob, @sunilUIET, @cmsbuild can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
Hi @aloeliger Few comments for now:
|
@srimanob Technically it will not crash because HLT has been simulating/providing their own non genuine GT inputs for a while now, but HLT hasn't attempted (to my knowledge) to officially move any inputs over to the GT emulator in preparation for this. i.e. I think you could run those steps, but I don't think you would be getting what you want, which is HLT run off of a GT emulator input for it's first step. I haven't actually tested this with the HLT step in place, but I guess I could add it if desired and see what we get. I think HLT experts would need to comment on what their plans are to provide upgrade changes to inputs for complete trigger workflows.
I have removed the |
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-43271/37680
|
Pull request #43271 was updated. @AdrianoDee, @srimanob, @cmsbuild, @sunilUIET, @miquork can you please check and sign again. |
Could you plese try to run the complete sequence, even HLT result is wrong: This will allow us to run the full workflow, including RECO and also DQM. It will open a chance to develop Phase-2 L1T DQM. I don't think we will compare HLT at this point, but (with this .78 workflow) it will also allow us to test if updated HLT code can run on L1P2GT. Thanks very much. |
@srimanob Trying to add the HLT step into this by itself just errors out, but adding it and DIGI2RAW raw step gets all steps functioning. I have added both into the workflow. |
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-43271/37691
|
@aloeliger Thx. |
FYI @SohamBhattacharya @rovere |
Well, for the menu studies we never used the FEVTDEBUG except for doing the I think our HLT experts @SohamBhattacharya and @rovere were making more use of that, though of course not using all the collections we have in the "full step-1 L1 menu". Given that we are working on Nano for the Phase-2 L1 Menu ntuples (https://github.com/cms-l1-dpg/Phase2-L1Nano) I think we could actually check whether we could produce these from the FEVTDEBUG. Would this be a good test? |
+1 Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-b34f90/35835/summary.html Comparison SummarySummary:
|
Hi @aloeliger |
+Upgrade New workflow with L1P2GT runs fine. |
@srimanob Yes. There aren't any plots to check, but at the least it will make sure everything still runs correctly, which I think is not a bad thing to test for. |
Kindly ping @cms-sw/pdmv-l2 as this PR will be useful for future Phase-2 L1T PR. Thanks. |
Ping again @cms-sw/pdmv-l2 |
+pdmv |
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. @rappoccio, @antoniovilela, @sextonkennedy (and backports should be raised in the release meeting by the corresponding L2) |
+1 |
PR description:
This PR creates an L1 workflow that includes the recently added Phase 2 L1 GT Emulator (and related
cmsDriver
step from #43210) for testing of L1 upgrade sequences and PRs. The workflow should run a modified version of theDigi
andDigiTrigger
steps with the GT emulator inserted, and it should be available on 2026 workflows. @srimanob This should provide the L1 testing workflow we have been discussing, please let me know if you would like/were envisioning something else.L1 would also like to explore whether this workflow, or a similar workflow idea can be used for validation of proposed to changes to L1, especially from the physics effects/menu effects perspective.
I don't know if this is the optimal solution for the testing that has been requested from L1, or if it will fully suit other L1 uses for it, so I would like some outside input on whether this is a good solution for being able to test the L1 steps going forward versus say, something like modifying the in place
Digi
/DigiTrigger
steps, and having these L1 steps be in place "by default". Also, this runs redundant steps from the L1 point of view... Anything past step2 is not L1, so as far as testing L1 exclusively, seems like wasted work, but I am not sure how I might change that, or if it is within my ability to do so, let alone whether it is a good idea.I don't know if there is a policy on naming/numbering these workflows. I didn't see anything relevant in the
README.md
, so I more or less chose.1001
at random. If this needs to be changed, no issue, it is not used in anything else that we plan to commit at the moment.@mcepeda & @artlbv I figure you might like to be informed about this, my hope is that eventually this can serve as the basis of a way to do more automatic L1 software validation and testing. Perhaps we can discuss this at some point.
PR validation:
This PR has been tested by running two added workflows via the commands:
runTheMatrix.py --what upgrade -l 20834.1001
for an older and more used geometryrunTheMatrix.py --what upgrade -l 26834.1001
for a more recent geometryBoth workflows manage to run through step2, with the
Digi
orDigiTrigger
steps, but, in full disclosure, step3 fails on the RAW2DIGI step, complaining about no FEDRawData collection.If this PR is a backport please specify the original PR and why you need to backport that PR. If this PR will be backported please specify to which release cycle the backport is meant for:
This PR is not a backport, and likely will not need backporting.