Update eigen, get omitted commit

[mrodozov]

See cms-externals/eigen-git-mirror#4

[cmsbuild]

A new Pull Request was created by @mrodozov (Mircho Rodozov) for branch IB/CMSSW_12_0_X/master.

@cmsbuild, @smuzaffar, @mrodozov, @iarspider can you please review it and eventually sign? Thanks.
cms-bot commands are listed here

[mrodozov]

please test

[mrodozov]

please test for slc7_aarch64_gcc9

[mrodozov]

please test for slc7_ppc64le_gcc9

[mrodozov]

please test for slc7_ppc64le_gcc9

[smuzaffar]

please test for CMSSW_12_0_X/slc7_ppc64le_gcc9

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15570/summary.html
COMMIT: 86fcbb8
CMSSW: CMSSW_12_0_X_2021-06-02-1100/slc7_amd64_gcc900
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmsdist/6971/15570/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:

• Triton: enable GPU support #6929 @kpedro88: Triton: enable GPU support

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15570/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15570/git-merge-result

Comparison Summary

Summary:

• No significant changes to the logs found
• Reco comparison results: 7 differences found in the comparisons
• DQMHistoTests: Total files compared: 37
• DQMHistoTests: Total histograms compared: 2650486
• DQMHistoTests: Total failures: 12
• DQMHistoTests: Total nulls: 1
• DQMHistoTests: Total successes: 2650451
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.004 KiB( 36 files compared)
• DQMHistoSizes: changed ( 312.0 ): 0.004 KiB MessageLogger/Warnings
• Checked 155 log files, 37 edm output root files, 37 DQM output files
• TriggerResults: no differences found

[cmsbuild]

-1

Failed Tests: UnitTests
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15572/summary.html
COMMIT: 86fcbb8
CMSSW: CMSSW_12_0_X_2021-06-02-1100/slc7_aarch64_gcc9
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmsdist/6971/15572/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:

• Triton: enable GPU support #6929 @kpedro88: Triton: enable GPU support

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15572/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15572/git-merge-result

Unit Tests

I found errors in the following unit tests:

---> test PhiTest had ERRORS

[cmsbuild]

-1

Failed Tests: UnitTests
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15583/summary.html
COMMIT: 86fcbb8
CMSSW: CMSSW_12_0_X_2021-06-02-2300/slc7_ppc64le_gcc9
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmsdist/6971/15583/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:

• [classlib] Fix needed for GCC 11 #6970 @cms-sw: [classlib] Fix needed for GCC 11

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15583/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15583/git-merge-result

Unit Tests

I found errors in the following unit tests:

---> test test_PrepareInputDb had ERRORS
---> test test_MpsWorkFlow had ERRORS
---> test DFM_Math had ERRORS
---> test DFM_MathRepr had ERRORS
and more ...

[mrodozov]

@ericcano @fwyzard one test failed on ARM would you see if there's anything wrong I can see the test ran and it's saying it used long double
we've changed this recently:
cms-sw/cmssw#33892
which includes Phi.h which includes deltaPhi and angle_units

[ericcano]

@mrodozov would it be possible to re-try (possibly against another IB)?

The logs for the test against CMSSW_12_0_X_2021-06-02-1100/slc7_aarch64_gcc9 are do not seem to matching the expected output from the code (there is no assertion in the area where the logs stop).

The logs CMSSW_12_0_X_2021-06-02-2300/slc7_ppc64le_gcc9 indicate memory allocation issues in CUDA and a files access error.

[mrodozov]

test parameters:

• release/arch = CMSSW_12_0_X_2021-06-03-2300/slc7_aarch64_gcc9

[mrodozov]

is last night IB ok ? it's a full build and only 3 unit test (unrelated) fail in it

[mrodozov]

please test

[ericcano]

@mrodozov, yes CMSSW_12_0_X_2021-06-03-2300 seems like a good baseline, thanks!

[cmsbuild]

-1

Failed Tests: UnitTests
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15651/summary.html
COMMIT: 86fcbb8
CMSSW: CMSSW_12_0_X_2021-06-03-2300/slc7_aarch64_gcc9
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmsdist/6971/15651/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:

• [BuildRules] Fix TSAN build issues #6985 @cms-sw: [BuildRules] Fix TSAN build issues

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15651/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-30262c/15651/git-merge-result

Unit Tests

I found errors in the following unit tests:

---> test PhiTest had ERRORS

[fwyzard]

Unit Tests

I found errors in the following unit tests:

---> test PhiTest had ERRORS

Well, looks like it is reproducible ...

[ericcano]

Indeed, the test exists at the same place (from the output). But the function which stops is not expected to stop at that point due to an assertion or test. It is a 3 step performance measurement test which never fails, and it stops after the 1st step. We should have a block of 3 runs like here:

Test 3 versions big decre
Fast version after 1234567980 iterations is 4.398231250149003
Time diff is  600.706042838
Simple version after 1234567980 iterations is 4.398231250149003
Time diff is  220.472635812
ReduceRange after 1234567980 iterations is 4.398231250149003
Time diff is  470.251683834

but test for big incr abruptly stops here:

Test 3 versions big incr
Fast version after 1234567980 iterations is 1.884954057030583
Time diff is  547.5369472

The last log is from this line: https://cmssdt.cern.ch/lxr/source/DataFormats/GeometryVector/test/PhiTest.cc?v=CMSSW_12_0_X_2021-06-03-2300#0100

[ericcano]

The test turned out to take more that 1 hour in the current circumstances/load of the ARM build machine, and we timeout individual tests after one hour.
Both base IB and IB running against the new Eigen ran in similar times when run by hand:

...
Test 3 versions big incr
Fast version after 1234567980 iterations is 1.884954057030583
Time diff is  558.000620182
Simple version after 1234567980 iterations is 1.884954057030583
Time diff is  174.290960233
ReduceRange after 1234567980 iterations is 1.884954057030583
Time diff is  470.337560569

real    68m40.245s
user    68m38.962s
sys     0m0.229s

and

...
Test 3 versions big incr
Fast version after 1234567980 iterations is 1.884954057030583
Time diff is  554.9811655460001
Simple version after 1234567980 iterations is 1.884954057030583
Time diff is  172.712609137
ReduceRange after 1234567980 iterations is 1.884954057030583
Time diff is  470.612648628

real    68m51.581s
user    68m45.160s
sys     0m1.780s

respectively.

A PR has been proposed to reduce the test length to seconds: cms-sw/cmssw#34029

[smuzaffar]

+externals
thanks @ericcano for checking this

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next IB/CMSSW_12_0_X/master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @silviodonato, @dpiparo, @qliphy (and backports should be raised in the release meeting by the corresponding L2)

[qliphy]

+1