[Buildrules] Treat ROCM_CXXFLAGS/LDFLAGS as valid compiler flags

[smuzaffar]

No description provided.

[cmsbuild]

A new Pull Request was created by @smuzaffar (Malik Shahzad Muzaffar) for branch IB/CMSSW_13_0_X/master.

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

[smuzaffar]

please test with cms-sw/cmssw#40619

[cmsbuild]

-1

Failed Tests: Build
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-aa10a1/30237/summary.html
COMMIT: a511b62
CMSSW: CMSSW_13_0_X_2023-01-29-1100/el8_amd64_gcc11
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmsdist/8271/30237/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:

• @missirol reduce verbosity of reHLT customisation in SKIM step cmssw#40599

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-aa10a1/30237/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-aa10a1/30237/git-merge-result

Build

I found compilation error when building:

>> Package HeterogeneousCore/ROCmServices built
>> Compiling  /data/cmsbld/jenkins/workspace/ib-run-pr-tests/CMSSW_13_0_X_2023-01-29-1100/src/HeterogeneousCore/ROCmServices/bin/rocmComputeCapabilities.cpp
>> Compiling  /data/cmsbld/jenkins/workspace/ib-run-pr-tests/CMSSW_13_0_X_2023-01-29-1100/src/HeterogeneousCore/ROCmServices/bin/isRocmDeviceSupported.hip.cc
>> Building binary rocmComputeCapabilities
/data/cmsbld/jenkins/workspace/ib-run-pr-tests/testBuildDir/el8_amd64_gcc11/external/gcc/11.2.1-f9b9dfdd886f71cd63f5538223d8f161/bin/../lib/gcc/x86_64-redhat-linux-gnu/11.2.1/../../../../x86_64-redhat-linux-gnu/bin/ld: tmp/el8_amd64_gcc11/src/HeterogeneousCore/ROCmServices/bin/rocmComputeCapabilities/isRocmDeviceSupported.hip.cc.o:(.hipFatBinSegment+0x8): undefined reference to `__hip_fatbin'
collect2: error: ld returned 1 exit status
>> Deleted: tmp/el8_amd64_gcc11/src/HeterogeneousCore/ROCmServices/bin/rocmComputeCapabilities/rocmComputeCapabilities
gmake: *** [tmp/el8_amd64_gcc11/src/HeterogeneousCore/ROCmServices/bin/rocmComputeCapabilities/rocmComputeCapabilities] Error 1
>> Compiling  /data/cmsbld/jenkins/workspace/ib-run-pr-tests/CMSSW_13_0_X_2023-01-29-1100/src/HeterogeneousCore/ROCmServices/bin/rocmIsEnabled.cpp
>> Compiling  /data/cmsbld/jenkins/workspace/ib-run-pr-tests/CMSSW_13_0_X_2023-01-29-1100/src/HeterogeneousCore/ROCmServices/bin/isRocmDeviceSupported.hip.cc
>> Building binary rocmIsEnabled


[smuzaffar]

test parameters:

• full_cmssw = true

[smuzaffar]

please test

[smuzaffar]

@fwyzard , any idea why we get isRocmDeviceSupported.hip.cc.o:(.hipFatBinSegment+0x8): undefined reference to '__hip_fatbin' collect2: error: ld returned 1 exit status ? I tried adding --hip-link to hipcc but still not success

[fwyzard]

I think it's because the final link is being done with g++ instead of hipcc.

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-aa10a1/30238/summary.html
COMMIT: a511b62
CMSSW: CMSSW_13_0_X_2023-01-29-1100/el8_amd64_gcc11
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmsdist/8271/30238/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:

• @missirol reduce verbosity of reHLT customisation in SKIM step cmssw#40599

You can see more details here:
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-aa10a1/30238/git-recent-commits.json
https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-aa10a1/30238/git-merge-result

Comparison Summary

Summary:

• You potentially removed 3040 lines from the logs
• Reco comparison results: 4 differences found in the comparisons
• DQMHistoTests: Total files compared: 49
• DQMHistoTests: Total histograms compared: 3555495
• DQMHistoTests: Total failures: 3
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3555470
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 48 files compared)
• Checked 211 log files, 162 edm output root files, 49 DQM output files
• TriggerResults: no differences found

[smuzaffar]

I think it's because the final link is being done with g++ instead of hipcc.

@fwyzard that is true and using hipcc forlinking needs a careful thinking and I will do not in next PR . By the way, I was able to link and run hip application if I explicitly link with rccl library (i.e I added <lib name="rccl"/> in rocm.xml). So can we still use g++ and link rccl?

[fwyzard]

No, that is not enough to make the program work.

$ g++ -O2 ... tmp/el8_amd64_gcc11/src/HeterogeneousCore/ROCmServices/bin/rocmComputeCapabilities/rocmComputeCapabilities.cpp.o tmp/el8_amd64_gcc11/src/HeterogeneousCore/ROCmServices/bin/rocmComputeCapabilities/isRocmDeviceSupported.hip.cc.o ... -lamdhip64 -lrccl -o tmp/el8_amd64_gcc11/src/HeterogeneousCore/ROCmServices/bin/rocmComputeCapabilities/rocmComputeCapabilities

does link, but the resulting binary does not run:

$ ./tmp/el8_amd64_gcc11/src/HeterogeneousCore/ROCmServices/bin/rocmComputeCapabilities/rocmComputeCapabilities
"Cannot find Symbol"
Aborted (core dumped)

I think because it is still missing the generation and linking of the device-side code and kernel launches.

[cmsbuild]

Pull request #8271 was updated.