Go to the Indico Home Page Oldest event Older event Up to category Newer event Newest event Printable version
  • Compact style
  • Indico style
  • Indico style - inline minutes
  • Indico style - numbered
  • Indico style - numbered + minutes
  • Indico Weeks View
    • Switch to the management area of this event
    Login

    EP R&D Software Working Group Meeting

    → Europe/Zurich
    32/1-A24 (CERN)

    32/1-A24

    CERN

    40
    Show room on map
    Live Notes
    EP R&D Software Working Group Meeting
    Zoom Meeting ID
    62893115044
    Host
    Graeme A Stewart
    Alternative host
    Andre Sailer
    Useful links
    Join via phone
    Zoom URL
    Hide

    Software R&D Working Meeting Minutes

    2022-06-15

    Apologies: Andi

    Room: Graeme, Andre, Erica, Anna, Dalila, Witek, Placido, Paul, Valentin, Javier

    Remote: Pere, Marco, Benedikt

    Introduction

    Reminder that there is an EP R&D Day BBQ on Tuesday, 10CHF (newdle link is at the bottom of the Indico agenda)

    Please upload your slides for the R&D presentaion ASAP (by tomorrow), so that people can comment on Friday before we upload.

    Reconstruction Tracking

    Graeme - Filters and fitters - is the basic space now covered?

    • Yes.

    Graeme - IRIS-HEP sponsored development of mkFit, which primarily targeted CMS. Could it be used/useful in ACTS?

    • Approximations and assumptions were rather CMS specific AFAICR; however, in principle, probably

    Marco Slide 17 - gain of x3 seems low, for a very parallel algorithm (CPU number is for a single core)

    • CCL is probably the problem here, it’s serial
    • Think about redesigning the algorithm, because the problem is local

    Anna - GNN - would be interesting to know the backends

    Erica - for validation in the CI, how many events?

    • 100k muons.
    • Would you do it also with electrons? For now it was more convenient with muons, but want to use electrons as well to be closer to real physics use cases.

    Valentin - is it a dedicated GPU?

    • Yes. It’s dedicated (and under Paul’s desk - CAD workstation).
    • Valentin didn’t find much documentation on how to do this.

    Andre - want to take a look at hist comparison code for key4hep.

    Erica - slide 12, what is the use case?

    • To formalise the output and have something more stable for the clients. Would also like to add dyanmic extra information (a la xAOD).

    Pere - magnetic field, template nesting is impressive!

    • Yes, it’s true and the idea is to resolve everything at compile time. Stephen is working on adding a local cell cache.
    • Interesting for AdePT; would like to know how accurate the texture memory interpolation is (Q. for Stephen!)
    There are minutes attached to this event. Show them.