Present/Contributing: Graeme Stewart, Pere Mato, Claire Antel, Eduardo Rodrigues, Krzysztof Genser, Efe Yazgan, Caterina Doglioni, Torre Wenaus, Patrick Gartung, Nicole Skidmore, Kyle Knoepfel, Nick Smith, Nicole Skidmore, Killian Lieret, Liz Sexton-Kennedy, David Lange, Valentin Volkl
Apologies/Contributing: Markus Diefenthaler, Benedikt Hegner
Vangelis Kourlitis presented a talk on HL-LHC and Beyond Computing Challenges. Thanks to him and to everyone who commented on the slides.
The pre-CHEP workshop had 150 people attening (not bad for a weekend!).
*Analysis Facilities:** presentation of what we have learned about AFs from the Analysis Facilities Forum. Key topics, such as AAI, DOMA, site “toolkit” (or “substrate”) as well as analyst needs were discussed. The AFF conveners released the first draft of their Analysis Facilities White Paper, which is still very much open for comments. See below for LHCC slides. Still a lot of open questions - very nice work by the forum conveners to bring all of this into the open for discussion. The White Paper will be informed by the workshop itself and the draft comments are being discussed and incorporated.
*Non-x86 Architectures:** presentations on power advantages of ARM (36% power savings on HEPScore workloads), along with the experiments’ experience of porting to ARM (ATLAS) and POWER9 (CMS), which is relatively strightforward. Progress in porting some middleware components and in running multi-arch jobs in the production systems - helped by increased availability at CERN and in the cloud. Some process is needed by which these resources can be pledged to be defined - this is something that will be discussed within WLCG (HSF can help). For GPU, 3/4 experiments using these at HLT in Run3, which improved efficiency, cost savings, etc (good talks from ALICE and CMS). Workloads do require significant tuning. No real experience of running mixed CPU/GPU workloads on WLCG - sites that have GPUs can help here, and we (HSF) can facilitate the process with the training. More discussion expected at the JENA computing workshop. There is a summary document available (which can be commented on!).
The next LHCC week is 5-9 June. For software, the referees asked us to report on the future role of Analysis Facilities. There will be a special 1 hour session on Monday 5 June as this is a big topic.
Alessandra Forti, one of the Analysis Facilities Forum co-conveners, will present: draft slides. Please comment on them by Monday 29 May, as they need to be posted for the referees by this date.
There will be some further dedicated discussion on the presentation at the Architects Forum meeting, Thursday 1 June at 14h CEST.
Only 17/22 projects with good students accepted - feedback from Google was that we had too many long projects. In previous years we had a 100% acceptance rate (there was also no difference in project size). For next year, we should balance long/short projects more.
Could have an HSF kick-off for all of the students? This would help foster a sense of community. Discuss with Benedikt / GSoC Coordinators / coordination list.
As mentioned above, there is a joint ECFA/AaPEC/NuPEC (JENA) meeting to discuss software and computing strategies for European science in Bologna, 12-14 June.
Nice papers at CHEP showing good progress in using Julia and a lot of potential. We’d like to organise more activities here, so there is a planning meeting next week, Tuesday 30 May, 16h CEST. Open to all!
The paper on Potential of the Julia programming language for high energy physics computing is now basically complete and ready for CSBS submission. The paper can be endorsed (there was a call on the
email@example.com list), but the deadline is today.
Change of topic for the next meeting - we will hear from the DOE funded HEP CCE project, which is now reaching the end of its first funding phase.
Meeting will be held jointly with ATLAS Core Software in Salle Curie, Wenesday 14 June at 16h30.
Benedikt started an issue in GitHub to gather ideas about reorganising and revamping the website to better reflect our areas of actual activity.
The next meeting will be on 8 June
Please sign up for chairing this or one of the future coordination meetings.