GPU computing & Machine Learning @ DESY

Europe/Berlin
SeminarRoom Flash (DESY Hamburg)

SeminarRoom Flash

DESY Hamburg

Building 28c, 2nd floor
Yves Kemp (DESY)
Description

several activities on GPU computing and machine learning are currently raising on and close to the DESY campus. Most of them have in common, that they use infrastructures in the DESY IT computing center, hence we (DESY IT) would like to

  •  learn from users what you need in terms of hardware and further support 
  • enable and strengthen inter-department communication and collaboration 
  • identify future common activities and follow-ups

We propose to have a round-table-style meeting on Tuesday, 6.11.2018, from 10:00-17:00 (max). We are pleased to announce that NVIDIA will be present on this day, potentially with a HPC architect with a science background. NVIDIA can show their hardware, software and training portfolio, and might also contribute to our discussions.

The agenda is entirely open for your contributions! The basic outline:

  •  All groups would present shortly an overview on their activities (10-15 minutes):
  • What problem do I need to solve using GPUs and/or machine learning techniques?
  • What is my application?
  • What is my current status?
  • Where can I get better?
  • Where do I need help?
  • Who could help me?
  • How?
  • Where can I help other people?

Please use the "apply here" button below to register! We need a rough number of participants for the free lunch and coffee, which will be available at the SeminarRoom.

Please note: the timetable is under construction & very tentative!

Please contact yves.kemp@desy.de and/or frank.schluenzen@desy.de for contributions!

    • Use Cases

      What are the GPU/Machine Learning based applications at the DESY group and institutes on campus? What are the problems and possible solutions?

      • 1
        Welcome and Introduction
        Speaker: Dr Yves Kemp (DESY)
        Slides
      • 2
        Use Case @ HZG
        Speaker: Dr Julian Moosmann (Institute of Materials Research, Helmholtz-Zentrum Geesthacht)
        Slides
      • 3
        Use Case ATLAS @ DESY
        Speaker: Paul Glaysher (DESY)
        Slides
      • 4
        Use Case Photon Science @ Petra3
        Speaker: Dr Xiaogang Yang (FS-Petra, DESY)
      • 5
        Use Case @ BELLE II
        Speaker: Torben Ferber (DESY)
        Slides
      • 11:17
        Coffee
      • 6
        GPU Image Processing for Near-Field Holography
        Speaker: Johannes Hagemann (DESY)
        Slides
      • 7
        Use Case Deep Learning @ DESY IT
        Speaker: Philipp Heuser (DESY)
        Slides
      • 8
        Use Case @ CFEL
        Speaker: Fabian Trost (CFEL)
        Slides
      • 9
        CryoEM & Software
        Speaker: Wolfgang Lugmayr (UKE)
        Slides
      • 10
        Use Case Ptychography @ Petra3
        Speaker: Dr Jan Garrevoet (DESY)
        Slides
      • 13:02
        Working Lunch & Discussions
      • 11
        Use Case CMS @ DESY
        Speaker: Mr Adam Elwood (DESY)
        Slides
      • 12
        Use Case Online Calibration @ Eur. XFEL
        Speaker: Steffen Hauf (Eur. XFEL)
        Slides
      • 13
        Laser Plasma Accelerators @ LUX
        Speakers: Manuel Kirchen (LUX), Sören Jalas (LUX)
      • 14
        Use Case CMS @Uni Hamburg
        Speaker: Gregor Kasieczka (Institut fuer Experimentalphysik / UHH)
        Slides
      • 15
        Roundup
        Speaker: Dr Yves Kemp (DESY)
        Slides
    • Nvidia GPUs

      GPU hardware, support, techniques, development

      • 16
        NVIDIA: Accelerated Computing: The path forward
        Speaker: Axel Koehler (NVIDIA)
        Slides