ECFA HL-LHC Trigger, Online, Offline Computing Prepatarory Group
================================================================

Meeting 2, 2014-2-29
--------------------

Indico Page:
https://indico.cern.ch/event/326755/

Present: STEWART, Graeme Andrew + SMITH, Wesley chair and minutes;
VANDE VYVRE, Pierre; MESCHI, Emilio; GLIGOROV, Vladimir; KRZEWICKI,
Mikolaj; PETERSEN, Brian; LANGE, David; GIRONE, Maria; STROM, David;
RIU DACHS, Imma; NEUFELD, Niko; KONSTANTINIDIS, Nikolaos; GORINI, Benedetto,
BOCCCI, Andrea

Apologies: CLARKE, Peter; MEIJERS, Frans

The meeting concentrated on the subjects to be addressed during the
topical workshop(s) during the week of 1-5 September.

After some discussion about overlaps between DAQ, Trigger and Offline,
and considering the ACAT conference as well, it was decided to:

- Move the offline workshop to the afternoon of Friday 5th

- Consider the workshop as a holistic event, so as to better address
  the strong overlaps between topics, so the topics guide the agenda
   items rather than break up the workshop

However, different focus areas will be addressed in the 3 sessions
which will be:

- DAQ, Thursday 2 PM  - 6 PM
- Trigger, Friday 9 AM - 1 PM
- Offline, Friday  2 PM - 6 PM

As some people will not be able to attend Thursday PM a very brief
summary should be given Friday AM.

Vava and Maria are soliciting rooms.

The workshops should provide the primary input material for the
Aix-Les-Bains workshop, but we also anticipate open questions that
will then need resoved between the workshop and the Aix-Les-Bains
meeting. 

In addition to the suggestions on the introductory slides the
following topics were discussed:

- For site issues we should distinguish between resources (money!) and
  how they are delivered. Ian Bird is the right person to propose a
  speaker for this topic.

- The model of doing primary reconstruction in the HLT has significant
  implications on timely calibration and data quality information
  availability.

- Simulation should be addressed, particularly with reference to
  associative memory devices (e.g., ATLAS FTK) and advanced L1 track
  triggers. These have been traditionally very hard to simulate and
  burn lots of CPU.

- We can consider how job submission systems might be unified between
  HLT and offline.

- Differences between full streaming experiments (ALICE, LHCb) and L1
  triggering (ATLAS, CMS) can be explored. When fully streaming
  decisions will be made _between_ signals as well as between signal
  and b/g.

- Technology review is very relevant to online as well: optical
  links, CPUs with FPGAs on die,  SERDES, network technology
  evolution.

In order to finalise the topics, and possible speakers, we should have
furthet discussions on the HL-LHC-ECFA-PG-TDOC mailing list (to avoid
surprising overlaps). However, PG members should be active in deciding
the topics for their specialist area. 

Next Meeting
------------

Should be in w/b July 7.