Software R&D Working Group Core Meeting

→ Europe/Zurich
32/S-C22 (CERN)

32/S-C22

CERN

17
Show room on map
Graeme Stewart (CERN), Jakob Blomer (CERN)
Description

This is a closed meeting for the Core/Editor group of the Software R&D Working Group, aimed at preparing the 2nd EP R&D Workshop.

Comments on slide

 

Slide 3: Adjust axis, CMS has updated numbers based on NanoAODs (LHCC Open Session), mention FCC-hh

Slide 4: Remove, wrong plot

Slide 6: It will take effort to exploit this silicon, "not a potential we reach today", HPC systems deploy these chips and we should use them

Slide 7: Lower latency of devices put pressure on software; key message: technology changes, software has to adapt

Slide 9: error --> uncertainty

Slide 11: Software for research vs research into software [combine with the graphics], sofware --> software

Slide 12: change IT liason, don't underline

 

Slide 16: show work packages, need to explore how best to do the training in the context of the overall workflow; "Integration" does not make sense for the reconstruction part; Integration --> Interfaces

 

Slide 19: graphics from ECFA slides: removal of the event loop

Slide 21: Are we chasing yesterday's technology? Answer: we decompose to make it more easy to adapt

Point one: doesn't allow to follow market driven development in order to keep the computing affordable

We don't want to use every technology but we want to explore be ready / in position to pick

Graphic: fastest way for a single person not necessarily the global optimum

 

Slide 22: remove flowgraph; "accumulate": implementation detail

Slide 23: Main point: move from static allocation to overprovisioning and dynamic provision / cooperate instead compete, graphics missing

Slide 24: Needs exchange between compute workflow, change "systems communicate with each other"

Flip slides 25 and 26: take a plot

Slide 25: Entry barrier is currently too high; inter-experiment barrier high because different terms / language is used; identify right building blocks, make connection to other R&D projects

Slide 27: a positive visual, remove last point

 

There are minutes attached to this event. Show them.
    • 11:00 → 13:00
      Discussion on EP R&D Workshop 2 Presentation 2h
      Speakers: Graeme Stewart (CERN), Jakob Blomer (CERN)