System performance modelling WG meeting

→ Europe/Zurich
31/S-028 (CERN)

31/S-028

CERN

30
Show room on map
  • Local: Andrea Sciabà, Catherine Biscarat
  • Remote: Johannes Elmsheuser, Michele Michelotto, David Lange, Oxana Smirnova

Andrea gives a summary of the recent meeting on site cost estimation.

Concerning consolidation, Markus argues that it is more about having a different model to operate "small sites" to save manpower. Catherine points out that e.g. in France sites get funding only if they provide storage (i.e. CPU-only sites are impossible).

About network cost calculations, Catherine says that these are already performed at IN2P3.

About sharing cost information among sites: in some cases it is impossible due to NDAs, but formulas can and should be shared.

Markus mentions his presentation at the WLCG MB about possible gains saying that it was well received. He also informs about the many summer students who have arrived, one of them studying the effect of caches on efficiency. A goal is to understand what is the optimal size for a cache.

Michele mentions Domenico's talk at the MB and the GDB: it is now widely accepted that a future HEP benchmark will be based on a combination of experiment workloads, and HS06 kept as a backup.

There are minutes attached to this event. Show them.