Refactor GPU clustering thresholds #608

czangela · 2021-03-10T08:41:49Z

PR description:

An effort to reduce dedicated namespaces and to introduce more consistency. see #32483 cms-sw

For more information see the commit names.

…hetics For more info see G.4.16.5. Constexpr variables in https://docs.nvidia.com/cuda/pdf/CUDA_C_Programming_Guide.pdf Previous commit was necessary for these changes, since std::min() takes arguments by reference

fwyzard · 2021-03-11T09:08:50Z

@czangela could you target the PRs to the CMSSW_11_3_X_Patatrack branch ?

The master branch is often unstable (currently it tracks CMSSW_11_3_0_pre4, while CMSSW_11_3_X_Patatrack is still on CMSSW_11_3_0_pre3).

I've changed the target for his one.

fwyzard · 2021-03-11T09:16:03Z

Validation summary

Reference release CMSSW_11_3_0_pre3 at 3220826
Development branch cms-patatrack/CMSSW_11_3_X_Patatrack at 0972c2f

🚧 Validation running at fu-c2a02-35-03:/data/user/fwyzard/patatrack/validation/run_608.c0ZqwRHPLW ...

czangela · 2021-03-11T09:21:26Z

@czangela could you target the PRs to the CMSSW_11_3_X_Patatrack branch ?

The master branch is often unstable (currently it tracks CMSSW_11_3_0_pre4, while CMSSW_11_3_X_Patatrack is still on CMSSW_11_3_0_pre3).

I've changed the target for his one.

Will keep this in mind, and thank you. :)

mmusich · 2021-03-11T10:15:39Z

HI @czangela @fwyzard
I have a question about the workflow. In principle these changes could be applied to cms-sw/cmssw as well now that PR cms-sw#31721 is merged. How do you wish to proceed for updates? Will someone from the patatrack team propose the changes to cms-sw/cmssw after they are validated here, or should we be doing the synching?
I see some other subsystems proposing changes directly to cms-sw/cmssw.
Thank you!

fwyzard · 2021-03-11T10:21:31Z

You are correct, but the problem is that some of the changes affect the pixel track and vertex reconstruction, which are not merged upstream yet.

Once they are tested in this branch, either Angela or myself can make a PR with only the pixel local reco part for cms-sw/cmssw.

Eventually, when the Patatrack branch is fully merged upstream, we can stop making PRs here.

mmusich · 2021-03-11T10:24:31Z

some of the changes affect the pixel track and vertex reconstruction

right (sigh). I overlooked that. This complicates also the validation procedure if we have to propose changes to algorithms... hopefully when it comes to that the tracking / vertexing part is merged.

czangela · 2021-03-17T08:40:46Z

Validation summary

Reference release CMSSW_11_3_0_pre3 at 3220826
Development branch cms-patatrack/CMSSW_11_3_X_Patatrack at 0972c2f

Validation plots

/RelValTTbar_14TeV/CMSSW_11_2_0-PU_112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

tracking validation plots and summary for workflow 11634.5
tracking validation plots and summary for workflow 11634.501
tracking validation plots and summary for workflow 11634.505
tracking validation plots and summary for workflow 11634.502
tracking validation plots and summary for workflow 11634.506

/RelValZMM_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

tracking validation plots and summary for workflow 11634.5
tracking validation plots and summary for workflow 11634.501
tracking validation plots and summary for workflow 11634.505
tracking validation plots and summary for workflow 11634.502
tracking validation plots and summary for workflow 11634.506

/RelValZEE_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v13-v1/GEN-SIM-DIGI-RAW

tracking validation plots and summary for workflow 11634.5
tracking validation plots and summary for workflow 11634.501
tracking validation plots and summary for workflow 11634.505
tracking validation plots and summary for workflow 11634.502
tracking validation plots and summary for workflow 11634.506

Validation plots (CPU vs GPU)

/RelValTTbar_14TeV/CMSSW_11_2_0-PU_112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

tracking validation plots and summary for workflows 11634.502 and 11634.501
tracking validation plots and summary for workflows 11634.506 and 11634.505

/RelValZMM_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

tracking validation plots and summary for workflows 11634.502 and 11634.501
tracking validation plots and summary for workflows 11634.506 and 11634.505

/RelValZEE_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v13-v1/GEN-SIM-DIGI-RAW

tracking validation plots and summary for workflows 11634.502 and 11634.501
tracking validation plots and summary for workflows 11634.506 and 11634.505

Throughput plots

/EphemeralHLTPhysics1/Run2018D-v1/RAW run=323775 lumi=53

logs and `nvprof`/`nvvp` profiles

/RelValTTbar_14TeV/CMSSW_11_2_0-PU_112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

reference release, workflow 11634.5
- ✔️ step3.py: log
development release, workflow 11634.5
- ✔️ step3.py: log
development release, workflow 11634.501
- ✔️ step3.py: log
development release, workflow 11634.505
- ✔️ step3.py: log
development release, workflow 11634.502
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 11634.506
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 11634.511
- ✔️ step3.py: log
development release, workflow 11634.512
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ❌ cuda-memcheck --tool synccheck (report, log) found no CUDA-MEMCHECK results
development release, workflow 11634.521
- ✔️ step3.py: log
development release, workflow 11634.522
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 136.885502
development release, workflow 136.885512
development release, workflow 136.885522

/RelValZMM_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

reference release, workflow 11634.5
- ✔️ step3.py: log
development release, workflow 11634.5
- ✔️ step3.py: log
development release, workflow 11634.501
- ✔️ step3.py: log
development release, workflow 11634.505
- ✔️ step3.py: log
development release, workflow 11634.502
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 11634.506
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 11634.511
- ✔️ step3.py: log
development release, workflow 11634.512
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ❌ cuda-memcheck --tool synccheck (report, log) found no CUDA-MEMCHECK results
development release, workflow 11634.521
- ✔️ step3.py: log
development release, workflow 11634.522
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 136.885502
development release, workflow 136.885512
development release, workflow 136.885522

/RelValZEE_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v13-v1/GEN-SIM-DIGI-RAW

reference release, workflow 11634.5
- ✔️ step3.py: log
development release, workflow 11634.5
- ✔️ step3.py: log
development release, workflow 11634.501
- ✔️ step3.py: log
development release, workflow 11634.505
- ✔️ step3.py: log
development release, workflow 11634.502
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 11634.506
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 11634.511
- ✔️ step3.py: log
development release, workflow 11634.512
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ❌ cuda-memcheck --tool synccheck (report, log) found no CUDA-MEMCHECK results
development release, workflow 11634.521
- ✔️ step3.py: log
development release, workflow 11634.522
- ✔️ step3.py: log
- ❌ profile.py: log
- ✔️ cuda-memcheck --tool initcheck (report, log) did not find any errors
- ✔️ cuda-memcheck --tool memcheck --leak-check full --report-api-errors all (report, log) did not find any errors
- ✔️ cuda-memcheck --tool synccheck (report, log) did not find any errors
development release, workflow 136.885502
development release, workflow 136.885512
development release, workflow 136.885522

Logs

The full log is available at https://aczirkos.web.cern.ch/aczirkos/validation/pulls/ea56bfaf068d2ab12b1c2149734e3f2f20dcf7eb/log

fwyzard · 2021-03-17T09:06:31Z

thanks @czangela - did the validation script run out of the box, or did it need any updates ?

czangela added 4 commits March 9, 2021 10:29

Merge namespaces in gpuClusteringConstants.h

5232bcb

Use CUDA Math API functions instead of standard library ones in kernels

c1be125

Change namespace gpuClustering to gpuClusteringConstants

7b93d9c

fwyzard changed the base branch from master to CMSSW_11_3_X_Patatrack March 11, 2021 09:09

fwyzard added Pixels Pixels-related developments enhancement labels Mar 11, 2021

czangela closed this Mar 11, 2021

czangela reopened this Mar 11, 2021

Fix RecoLocalTracker/SiPixelClusterizer build file

98e0ec2

czangela closed this Mar 22, 2021

czangela mentioned this pull request Mar 26, 2021

NVProfilerService assertion failed at preModuleBeginStream cms-sw/cmssw#33278

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Refactor GPU clustering thresholds #608

Refactor GPU clustering thresholds #608

czangela commented Mar 10, 2021 •

edited

Loading

fwyzard commented Mar 11, 2021 •

edited

Loading

fwyzard commented Mar 11, 2021 •

edited

Loading

czangela commented Mar 11, 2021

mmusich commented Mar 11, 2021

fwyzard commented Mar 11, 2021

mmusich commented Mar 11, 2021

czangela commented Mar 17, 2021

fwyzard commented Mar 17, 2021

Refactor GPU clustering thresholds #608

Refactor GPU clustering thresholds #608

Conversation

czangela commented Mar 10, 2021 • edited Loading

PR description:

fwyzard commented Mar 11, 2021 • edited Loading

fwyzard commented Mar 11, 2021 • edited Loading

Validation summary

czangela commented Mar 11, 2021

mmusich commented Mar 11, 2021

fwyzard commented Mar 11, 2021

mmusich commented Mar 11, 2021

czangela commented Mar 17, 2021

Validation summary

Validation plots

/RelValTTbar_14TeV/CMSSW_11_2_0-PU_112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

/RelValZMM_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

/RelValZEE_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v13-v1/GEN-SIM-DIGI-RAW

Validation plots (CPU vs GPU)

/RelValTTbar_14TeV/CMSSW_11_2_0-PU_112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

/RelValZMM_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

/RelValZEE_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v13-v1/GEN-SIM-DIGI-RAW

Throughput plots

/EphemeralHLTPhysics1/Run2018D-v1/RAW run=323775 lumi=53

logs and nvprof/nvvp profiles

/RelValTTbar_14TeV/CMSSW_11_2_0-PU_112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

/RelValZMM_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v14-v1/GEN-SIM-DIGI-RAW

/RelValZEE_14/CMSSW_11_2_0-112X_mcRun3_2021_realistic_v13-v1/GEN-SIM-DIGI-RAW

Logs

fwyzard commented Mar 17, 2021

czangela commented Mar 10, 2021 •

edited

Loading

fwyzard commented Mar 11, 2021 •

edited

Loading

fwyzard commented Mar 11, 2021 •

edited

Loading

logs and `nvprof`/`nvvp` profiles