System performance modelling WG meeting
→
Europe/Zurich
600/R-001 (CERN)
- Local: Andrea Sciabà, Markus Schulz
- Remote: Michele Michelotto, Oxana Smirnova, Shigeki Misawa
Discussion on HPC and GPU resources. Currently, they are made available in a "bursty" fashion (for a limited time window) and this negatively impacts their usability.
We will have to prepare a proposal for the October RRB and before that, a report for September.
We will need to use workloads as benchmarks.
CMS is working on a framework for using GPUs for multiple workloads.
We will need to form a dedicated team in the benchmarking working group. Not realistic to have a benchmark by September.
There are minutes attached to this event.
Show them.