Summer Student Milla Bramsted

• She is working on benchmarking SoA code on GPUs
• We will track her project in this google doc
• She is adding CUDA kernels to this repo
• A first Kernel was integrated to the benchmarking framework
• Next steps:
  • Add more kernels
  • Add the other NGT SoA approaches again (were temporarily removed)
  • Re-organize the code

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)
• Next steps:
  • Run on the different GPUs that are now available, see ngt-resources for details:
    • • Nvidia: H100 188GB NVL (12x), H100 80GB SXM (6x), L40S 48GB (7x)
      • AMD: Instinct MI300X 192GB (2x), Radeon Pro W7900 48GB (6x)
  • Get the full CI to work
• Remark: Most NGT GPUs are busy these days

Implement NGT SoA Code in O2 standalone benchmark

• Working on this fork of the AliceO2 repo
• Simplified and optimized the SoA code in the last few weeks
• Everything is running and we have identified the classes apply our SoA code to
• Next steps:
  • Implement the SoA code on those classes