Change the default optimisation level to -O3

[fwyzard]

No description provided.

[cmsbuild]

A new Pull Request was created by @fwyzard for branch IB/CMSSW_14_1_X/master.

@smuzaffar, @iarspider, @cmsbuild, @aandvalenzuela can you please review it and eventually sign? Thanks.
@sextonkennedy, @rappoccio, @antoniovilela you are the release manager for this.
cms-bot commands are listed here

[cmsbuild]

cms-bot internal usage

[fwyzard]

cms-sw/cmssw#44931 shows a 2.4% speed up to the HLT, running with the x86-64-v3 microarch.

[fwyzard]

please test

[smuzaffar]

test parameters:

• full_cmssw = true
• enable = threading

[smuzaffar]

please test

lets re-build full cmssw

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-be9632/39335/summary.html
COMMIT: dd72101
CMSSW: CMSSW_14_1_X_2024-05-09-2300/el8_amd64_gcc12
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmsdist/9183/39335/install.sh to create a dev area with all the needed externals and cmssw changes.

The following merge commits were also included on top of IB + this PR after doing git cms-merge-topic:

• @Duchstf L1 Modify Btag Jet Types [Merged with 14_0_0_pre3] cmssw#44577
• @mmusich Pixel GPU client: fix siPixelPhase1RawDataErrorComparator collections post-alpaka migration cmssw#44933
• @p2l1-gtEmulator PuppiNNTau to Phase2GT interface refactor cmssw#44737
• @iarspider [GCC13] Fix Woverloaded-virtual warning cmssw#44581
• @DenkMybu Computation of silicon strip hit efficiency : Consecutive missing hits recovery cmssw#44858
• @Dr15Jones Added noexcept to methods called by async functions cmssw#44902
• @abhih1 Change parameters for ECAL MLDQM  cmssw#44929
• @makortel Use standard edm::exceptionContext() in Worker::prePrefetchSelectionAsync() cmssw#44944
• @BenjaminRS L1TP2: Bug Fix: Adding a setting of TTTrack word to allow Track Quality MVA to read variables cmssw#44897
• @jonamotta Fix L1NNCaloTau verbosity cmssw#44770

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-be9632/39335/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-be9632/39335/git-merge-result

Comparison Summary

Summary:

• You potentially removed 317 lines from the logs
• Reco comparison results: 12 differences found in the comparisons
• DQMHistoTests: Total files compared: 48
• DQMHistoTests: Total histograms compared: 3335848
• DQMHistoTests: Total failures: 17
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3335811
• DQMHistoTests: Total skipped: 20
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 47 files compared)
• Checked 202 log files, 165 edm output root files, 48 DQM output files
• TriggerResults: no differences found

[slava77]

Reco comparison results: 12 differences found in the comparisons

this doesn't have the usual numerical precision level differences.
The log looks like the build is full. Oddly clean.

[makortel]

Reco comparison results: 12 differences found in the comparisons

this doesn't have the usual numerical precision level differences.

What do you mean? We shouldn't be having random numerical-precision level differences that would occur "reliably" in CPU PR tests (as opposed to the "known random differences"), or have I missed something?

[slava77]

Reco comparison results: 12 differences found in the comparisons

this doesn't have the usual numerical precision level differences.

What do you mean? We shouldn't be having random numerical-precision level differences that would occur "reliably" in CPU PR tests (as opposed to the "known random differences"), or have I missed something?

a recompilation with different flags related to optimization typically leads to some bitwise differences. It's not the case here. I haven't looked closely enough at what -O3 does compared to the baseline/default.

[fwyzard]

On top of -O2, -O3 enables

• -fgcse-after-reload
• -fipa-cp-clone
• -floop-interchange
• -floop-unroll-and-jam
• -fpeel-loops
• -fpredictive-commoning
• -fsplit-loops
• -fsplit-paths
• -ftree-loop-distribution
• -ftree-partial-pre
• -funroll-completely-grow-size (undocumented)
• -funswitch-loops
• -fvect-cost-model=dynamic
• -fversion-loops-for-strides

According to the GCC documentation these should only deal with invariant changes. So, unlike with -Ofast, -ffast-math and -ffp-contract, I think it is expected that there aren't any differences in the results.

[smuzaffar]

please test

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-be9632/39349/summary.html
COMMIT: dd72101
CMSSW: CMSSW_14_1_X_2024-05-12-0000/el8_amd64_gcc12
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmsdist/9183/39349/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• You potentially removed 151 lines from the logs
• Reco comparison results: 2 differences found in the comparisons
• DQMHistoTests: Total files compared: 48
• DQMHistoTests: Total histograms compared: 3335848
• DQMHistoTests: Total failures: 0
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3335828
• DQMHistoTests: Total skipped: 20
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 47 files compared)
• Checked 202 log files, 165 edm output root files, 48 DQM output files
• TriggerResults: no differences found

[smuzaffar]

please test for el9_amd64_gcc12

[smuzaffar]

please test for el8_aarch64_gcc12

[smuzaffar]

please test for el8_amd64_gcc12

lets test with thread enabled too

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-be9632/39448/summary.html
COMMIT: dd72101
CMSSW: CMSSW_14_1_X_2024-05-20-2300/el8_aarch64_gcc12
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmsdist/9183/39448/install.sh to create a dev area with all the needed externals and cmssw changes.

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-be9632/39449/summary.html
COMMIT: dd72101
CMSSW: CMSSW_14_1_X_2024-05-20-2300/el8_amd64_gcc12
Additional Tests: THREADING
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmsdist/9183/39449/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• You potentially removed 141 lines from the logs
• Reco comparison results: 17 differences found in the comparisons
• DQMHistoTests: Total files compared: 48
• DQMHistoTests: Total histograms compared: 3338862
• DQMHistoTests: Total failures: 854
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3337988
• DQMHistoTests: Total skipped: 20
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 47 files compared)
• Checked 202 log files, 165 edm output root files, 48 DQM output files
• TriggerResults: no differences found

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-be9632/39447/summary.html
COMMIT: dd72101
CMSSW: CMSSW_14_1_X_2024-05-20-2300/el9_amd64_gcc12
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmsdist/9183/39447/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• You potentially added 70 lines to the logs
• Reco comparison results: 56417 differences found in the comparisons
• DQMHistoTests: Total files compared: 48
• DQMHistoTests: Total histograms compared: 3338862
• DQMHistoTests: Total failures: 146340
• DQMHistoTests: Total nulls: 282
• DQMHistoTests: Total successes: 3192220
• DQMHistoTests: Total skipped: 20
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: -0.273 KiB( 47 files compared)
• DQMHistoSizes: changed ( 10224.0 ): -0.903 KiB SiStrip/MechanicalView
• DQMHistoSizes: changed ( 13034.0 ): 0.862 KiB SiStrip/MechanicalView
• DQMHistoSizes: changed ( 141.044 ): -0.012 KiB JetMET/SUSYDQM
• DQMHistoSizes: changed ( 141.046 ): 0.004 KiB JetMET/SUSYDQM
• DQMHistoSizes: changed ( 250202.181 ): -0.288 KiB SiStrip/MechanicalView
• DQMHistoSizes: changed ( 25202.0 ): 0.064 KiB SiStrip/MechanicalView
• Checked 202 log files, 165 edm output root files, 48 DQM output files

[smuzaffar]

+externals
lets get this in GEANT4 IBs first

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next IB/CMSSW_14_1_X/g4 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)