Update CUDA to version 10.1.105, on x86_64

[fwyzard]

• CUDA version 10.1.105
  Support for gcc 8 and clang 7.
  Include new Matrix Multiply library cublasLt.
  Various fxes to CUDA libraries and NVCC compiler.

• NVIDIA drivers version 418.39

See https://docs.nvidia.com/cuda/archive/10.1/cuda-toolkit-release-notes/index.html .

[cmsbuild]

A new Pull Request was created by @fwyzard (Andrea Bocci) for branch IB/CMSSW_10_6_X/gcc700.

@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]

@smuzaffar could you check if the GPU-equipped machines at CERN are compatible with CUDA 10.1 ?

We have updated all the local machines we have access to, and the ones at Flatiron are being updated as well.

[fwyzard]

@cmsbuild please test

[cmsbuild]

The tests are being triggered in jenkins.
https://cmssdt.cern.ch/jenkins/job/ib-any-integration/34223/console

[fwyzard]

This should bring CUDA for the gcc 7 builds back in sync with the gcc 8 builds.

[cmsbuild]

+1
Tested at: ec2ad19
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-4886/34223/summary.html

[cmsbuild]

Comparison job queued.

[cmsbuild]

Comparison is ready
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-4886/34223/summary.html

Comparison Summary:

• No significant changes to the logs found
• Reco comparison results: 0 differences found in the comparisons
• DQMHistoTests: Total files compared: 32
• DQMHistoTests: Total histograms compared: 3142783
• DQMHistoTests: Total failures: 1
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3142585
• DQMHistoTests: Total skipped: 197
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 31 files compared)
• Checked 133 log files, 14 edm output root files, 32 DQM output files

[fabiocos]

+1

[fabiocos]

@mrodozov I will move forward with the merge, @fwyzard verified the CUDA version used on the CERN machine used for tests

[fabiocos]

merge