Improve Eigen compatibility with CUDA (10.4.x devel)

[fwyzard]

Update Eigen to the master branch as of Tue Sep 25 20:26:16 2018 +0200

• hg hash 66ba78bf7efa93f69a075830a87a010ed1b1fe30
• git hash 01ae86b9aad30b1e65cf1b749fd6cd9a645ac00d

Patch Tensorflow to follow Eigen internal changes

• cherry-pick changes from upstream repository
• add local changes for the latest updates

Improve Eigen compatibility with CUDA

• add support for cache-size queries on CUDA devices.
• extend support for matrix inversion on CUDA devices above 4x4 matrices;
  the size of the matrices that can be inverted is limited at runtime by the per-thread stack size.
• extend support for diagonal matrices on CUDA devices.
• fix deprecation warning in CUDA 10.0.

[cmsbuild]

A new Pull Request was created by @fwyzard (Andrea Bocci) for branch IB/CMSSW_10_3_X/gcc700next.

@cmsbuild, @smuzaffar, @gudrutis, @mrodozov can you please review it and eventually sign? Thanks.
You can sign-off by replying to this message having '+1' in the first line of your reply.
You can reject by replying to this message having '-1' in the first line of your reply.

[fwyzard]

@cmsbuild, please test

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/30687/console

[cmsbuild]

Pull request #4381 was updated.

[fwyzard]

@cmsbuild, please test

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/30691/console

[cmsbuild]

Comparison not run due to runTheMatrix errors (RelVals and Igprof tests were also skipped)

[smuzaffar]

please test

[cmsbuild]

Comparison not run due to runTheMatrix errors (RelVals and Igprof tests were also skipped)

[fwyzard]

@cmsbuild please test

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/31461/console

[cmsbuild]

Comparison not run due to runTheMatrix errors (RelVals and Igprof tests were also skipped)

[fwyzard]

Do you have any idea why this fails, while the same changes (#4474) work on the standard IB ?

[davidlange6]

please test

…

On Nov 8, 2018, at 2:09 PM, Andrea Bocci ***@***.***> wrote: Do you have any idea why this fails, while the same changes (#4474) work on the standard IB ? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[davidlange6]

the error doesn't look related to eigen. lets try again

…

On Nov 8, 2018, at 2:09 PM, Andrea Bocci ***@***.***> wrote: Do you have any idea why this fails, while the same changes (#4474) work on the standard IB ? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/31533/console

[cmsbuild]

+1
Tested at: 8ca4f41
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-4381/31533/summary.html

[cmsbuild]

Comparison job queued.

[cmsbuild]

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-4381/31533/summary.html

@slava77 comparisons for the following workflows were not done due to missing matrix map:

• /build/cmsbld/jenkins/workspace/compare-root-files-short-matrix/results/JR-comparison/PR-4381/7.3_CosmicsSPLoose_UP18+CosmicsSPLoose_UP18+DIGICOS_UP18+RECOCOS_UP18+ALCACOS_UP18+HARVESTCOS_UP18

Comparison Summary:

• No significant changes to the logs found
• Reco comparison results: 4 differences found in the comparisons
• DQMHistoTests: Total files compared: 32
• DQMHistoTests: Total histograms compared: 3013827
• DQMHistoTests: Total failures: 1
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3013629
• DQMHistoTests: Total skipped: 197
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 31 files compared)
• Checked 134 log files, 14 edm output root files, 32 DQM output files

[smuzaffar]

lets get this in devel IBs and after 10.4.0.pre2 we can merged include it in normal IBs.

[fwyzard]

Thanks.

I'd rather have it in pre2 than after, but I can bring it up at the next ORP (unless pre2 srrives before that).

[smuzaffar]

@fabiocos wants to build pre2 over the weekend.

[fwyzard]

I see... I'd still be happy to have this in (it saves me from building the externals myself for Patatrack) but I won't complain if it enters after pre2...