Dynamic Memory Allocation for LST Objects

[GNiendorf]

Dynamic Memory Limits

I compute a dynamic max for each module that is the max possible number of objects that can be created given the number of lower level objects (so for example the number of possible T5's for a given module given the total number of T3's in modules where two T3's can be connected). This upper bound provides a large reduction in the occupancies. This is also accomplished with no new truncations, since this dynamic max represents the maximum number of possible objects that can be created.

For the segments, I had to rearrange the addPixelSegmentToEvent() function so that the segments container is instantiated after the MDs container is filled. Right now the segments object is instantiated during the pLS loading, which is done before the MDs container is filled.

Object Size Reduction

Reduction in object sizes (total occupancy basically) at current max occupancies (averaged over 10 events, 0.8 GeV):

• 682,404 -> 303,891 (55% decrease) for MD's.
• 1,930,994->857,880 (56% decrease) for Segments.
• 1,980,235 -> 737,502 (63% decrease) for T3's.
• 502,053 -> 212,953 (58% decrease) for T5's.

Total Memory Decrease

For single stream, 100 events at current occupancies:

1695MiB -> 1275MiB (25% decrease, 0.8 GeV)

The code has been tested locally: it compiles and gives the identical results (on CPU) for LST.

[cmsbuild]

cms-bot internal usage

[slava77]

test parameters:

• enable_tests = gpu
• workflows_gpu = 29634.704,29834.704
• workflows = 29634.703,29834.703
• relvals_opt = -w upgrade,standard
• relvals_opt_gpu = -w upgrade,standard

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-47232/43512

[cmsbuild]

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

It involves the following packages:

• RecoTracker/LSTCore (reconstruction)

@cmsbuild, @jfernan2, @mandrenguyen can you please review it and eventually sign? Thanks.
@GiacomoSguazzoni, @VinInn, @VourMa, @dgulhan, @felicepantaleo, @gpetruc, @missirol, @mmusich, @mtosi, @rovere this is something you requested to watch as well.
@antoniovilela, @mandrenguyen, @rappoccio, @sextonkennedy you are the release manager for this.

cms-bot commands are listed here

[slava77]

@cmsbuild please test

[cmsbuild]

+1

Size: This PR adds an extra 116KB to repository
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-3a0d49/44114/summary.html
COMMIT: 2967c14
CMSSW: CMSSW_15_0_X_2025-01-31-1100/el8_amd64_gcc12
Additional Tests: GPU
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmssw/47232/44114/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• You potentially removed 1 lines from the logs
• Reco comparison results: 6 differences found in the comparisons
• DQMHistoTests: Total files compared: 53
• DQMHistoTests: Total histograms compared: 4194999
• DQMHistoTests: Total failures: 65
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 4194914
• DQMHistoTests: Total skipped: 20
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 52 files compared)
• Checked 230 log files, 200 edm output root files, 53 DQM output files
• TriggerResults: no differences found

GPU Comparison Summary

Summary:

• You potentially added 1 lines to the logs
• Reco comparison results: 171 differences found in the comparisons
• DQMHistoTests: Total files compared: 9
• DQMHistoTests: Total histograms compared: 117389
• DQMHistoTests: Total failures: 2301
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 115088
• DQMHistoTests: Total skipped: 0
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 8 files compared)
• Checked 32 log files, 36 edm output root files, 9 DQM output files
• TriggerResults: no differences found

[jfernan2]

+1

[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. @rappoccio, @mandrenguyen, @sextonkennedy, @antoniovilela (and backports should be raised in the release meeting by the corresponding L2)

[mandrenguyen]

+1