NextGenTrigger Task 1.7 Meeting

Europe/Zurich
40/R-D10 (CERN)

40/R-D10

CERN

20
Show room on map
Zoom Meeting ID
65606242618
Host
Axel Naumann
Alternative hosts
Andrea Bocci, Attila Krasznahorkay
Useful links
Join via phone
Zoom URL
    • 09:30 09:35
      Project updates 5m
      Speakers: Dr Andrea Bocci (CERN), Attila Krasznahorkay (University of Massachusetts (US)), Axel Naumann (CERN)
    • 09:40 10:20
      Reports and updates 40m
      Speaker: All

      Eric:

      Investigation on CUDA launch strategies showed that calling CUDA runtime from a single thread is a must for performance. Kernel launch time is 5x longer with multiple threads and host function launch even 100x longer. I'll prepare the full presentation for next week,

       

      Oliver:

      Summer Student

      • ALICE summer student has arrived: Milla Bramsted
      • She is working on benchmarking SoA code on GPUs.
      • We will track her project in this google doc.
      • She will add CUDA kernels to this repo.

      ALICE O2 CI-Pipelines on NGT Cluster

      • A fork of the AliceO2 repo is not in the  NextGenTrigggers (NGT) GitHub organization.
      • It has a GitHub action running the standalone benchmark on NGT GPUs (H100).
      • Uses the builds in /cvmfs/alice.cern.ch/ of O2 and dependencies (pipeline takes about 7 minutes).

       

    • 10:25 10:30
      AOB 5m