Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[WIP] Follow-ups to the CUDA framework PR review #429

Merged

Conversation

makortel
Copy link

@makortel makortel commented Dec 5, 2019

PR description:

This PR is to collect the follow-ups to the review of cms-sw#28537. Work in progress until cms-sw#28537 gets merged.

PR validation:

Code compiles, unit tests run, profiling workflow runs.

@makortel
Copy link
Author

@fwyzard Would it be possible to run the validation with this PR now to check the impact on performance by the last two commits? (my test on felk40 with 8 EDM streams/threads shows little to no effect)

@makortel
Copy link
Author

(fixed the code format)

@fwyzard
Copy link

fwyzard commented Dec 17, 2019

Validation summary

Reference release CMSSW_11_0_0_pre13 at 91be707
Development branch cms-patatrack/CMSSW_11_0_X_Patatrack at e41560b

🚧 Validation running at fu-c2a02-37-01:/data/user/fwyzard/patatrack/validation/run.rv8HiWsATC ...

@fwyzard
Copy link

fwyzard commented Dec 17, 2019

sure - though I suspect the throughput scan is broken, but at least I can look into what's wrong with it ...

@makortel
Copy link
Author

sure -

Thanks!

though I suspect the throughput scan is broken,

:(

but at least I can look into what's wrong with it ...

:)

@fwyzard
Copy link

fwyzard commented Dec 18, 2019

Mhm, the validation failed during git cms-checkdeps -a after the PR was merged locally...
I am trying again by hand.

@makortel
Copy link
Author

The changes in WaitingTaskWithArenaHolder.h may cause the universe to be checked out (IIRC I got bitten that as well at some point). The last two commits should be independent from the others. If it helps, I can cherry-pick them to another branch/PR.

@fwyzard
Copy link

fwyzard commented Dec 18, 2019

No, there seems to be a more general issue with the validation tools and/or machines, but it's hard to pinpoint because the tests seem to work fine when run by hand.

@AdrianoDee
Copy link

🚧 Validation running at fu-c2d41-12-01:/data/user/adiflori/patatrack-validation/run.x9gf7Uxgfw

@AdrianoDee
Copy link

Validation summary

Reference release CMSSW_11_0_0_pre13 at 91be707
Development branch cms-patatrack/CMSSW_11_0_X_Patatrack at e41560b
Testing PRs:

Validation plots

/RelValTTbar_13/CMSSW_10_6_0-PU25ns_106X_upgrade2018_realistic_v4-v1/GEN-SIM-DIGI-RAW

  • tracking validation plots and summary for workflow 10824.5
  • tracking validation plots and summary for workflow 10824.501
  • tracking validation plots and summary for workflow 10824.502

/RelValZMM_13/CMSSW_10_6_0-PU25ns_106X_upgrade2018_realistic_v4-v1/GEN-SIM-DIGI-RAW

  • tracking validation plots and summary for workflow 10824.5
  • tracking validation plots and summary for workflow 10824.501
  • tracking validation plots and summary for workflow 10824.502

/RelValTTbar_13/CMSSW_10_6_0-PU25ns_106X_upgrade2018_design_v3-v1/GEN-SIM-DIGI-RAW

  • tracking validation plots and summary for workflow 10824.5
  • tracking validation plots and summary for workflow 10824.501
  • tracking validation plots and summary for workflow 10824.502

Throughput plots

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

scan-136.885502.png
zoom-136.885502.png

logs and nvprof/nvvp profiles

/RelValTTbar_13/CMSSW_10_6_0-PU25ns_106X_upgrade2018_realistic_v4-v1/GEN-SIM-DIGI-RAW

  • reference release, workflow 10824.5
  • development release, workflow 10824.5
  • development release, workflow 10824.501
  • development release, workflow 10824.502
    • ✔️ 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 136.885502
  • testing release, workflow 10824.5
  • testing release, workflow 10824.501
  • testing release, workflow 10824.502
    • ✔️ 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
  • testing release, workflow 136.885502

/RelValZMM_13/CMSSW_10_6_0-PU25ns_106X_upgrade2018_realistic_v4-v1/GEN-SIM-DIGI-RAW

  • reference release, workflow 10824.5
  • development release, workflow 10824.5
  • development release, workflow 10824.501
  • development release, workflow 10824.502
    • ✔️ 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 136.885502
  • testing release, workflow 10824.5
  • testing release, workflow 10824.501
  • testing release, workflow 10824.502
    • ✔️ 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
  • testing release, workflow 136.885502

/RelValTTbar_13/CMSSW_10_6_0-PU25ns_106X_upgrade2018_design_v3-v1/GEN-SIM-DIGI-RAW

  • reference release, workflow 10824.5
  • development release, workflow 10824.5
  • development release, workflow 10824.501
  • development release, workflow 10824.502
    • ✔️ 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 136.885502
  • testing release, workflow 10824.5
  • testing release, workflow 10824.501
  • testing release, workflow 10824.502
    • ✔️ 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
  • testing release, workflow 136.885502

Logs

The full log is available at https://patatrack.web.cern.ch/patatrack/validation/pulls/8e22bf5430838b6e504de51b457a5665f4fd3bca/log .

@makortel
Copy link
Author

Thanks, so indeed no effect on the throughput.

- Remove obsolete comment
- Make computeCapability() const
- Remove unnecessary ActivityRegistry argument from constructor
- Replace C-style cast with static_cast
- Remove test_main.cc as unnecessary
- explicit CUDAThing constructor
- Make EDModule member variables const when possible
- Add descriptions.setComment() where one was missing
- Use EDPutToken and ctx.emplace() in TestCUDAProducerGPUFirst
- Simplify test config file and the runner script
…AAnalyzerGPU

The point is that the member functions mutating the visible state
should not be const even if they are thread safe.
TestCUDAAnalyzerGPUKernel::analyzeAsync() mutates the visible state
(by "filling a histogram"), so it should not be const. Declaring
TestCUDAAnalyzerGPU::gpuAlgo_ as mutable is an improvement since
filling the histogram does not really change the visible state of the
EDAnalyzer (towards the framework).
…ecking in CUDAScopedContextProduce destructor
…evices

Also add comment that requireDevices() is meant for unit tests only
fwyzard pushed a commit that referenced this pull request Oct 23, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 6, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 6, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 6, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 6, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 6, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 9, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 12, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 16, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 16, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 16, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Nov 16, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard added a commit that referenced this pull request Nov 27, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard added a commit that referenced this pull request Nov 27, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard added a commit that referenced this pull request Nov 28, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Dec 25, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Dec 25, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Dec 26, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard added a commit that referenced this pull request Dec 26, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard added a commit that referenced this pull request Dec 26, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Dec 29, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Dec 29, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Dec 29, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Dec 29, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Dec 29, 2020
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Jan 13, 2021
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Jan 15, 2021
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Jan 15, 2021
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Mar 23, 2021
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
fwyzard pushed a commit that referenced this pull request Apr 1, 2021
Rename the cudautils namespace to cms::cuda or cms::cudatest, and drop the CUDA prefix from the symbols defined there.

Always record and query the CUDA event, to minimize need for error checking in CUDAScopedContextProduce destructor.

Add comments to highlight the pieces in CachingDeviceAllocator that have been changed wrt. cub.

Various other updates and clean up:
  - enable CUDA for compute capability 3.5.
  - clean up CUDAService, CUDA tests and plugins.
  - add CUDA existence protections to BuildFiles.
  - mark thread-safe static variables with CMS_THREAD_SAFE.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants