Collaboration Meeting

Europe/Berlin
Hörsaal - R 140 (Steinbuch Centre for Computing, Building 449, Campus North)

Hörsaal - R 140

Steinbuch Centre for Computing, Building 449, Campus North

Karlsruhe Institute of Technology - Campus North Hermann-von-Helmholtz-Platz 1 76344 Eggenstein-Leopoldshafen

Introduction by Thomas Kuhr

  • The logo should be on the web page for download (done after the meeting).
  • We decided to collect talks, proceedings, papers on the web page. Please send them to nikolai.hartmann@physik.uni-muenchen.de.
  • It was suggested to hold the next collaboration meeting in Bonn during the DPG conference. The Bonn colleagues will follow up on this idea.
  • We should improve the interaction among groups by having more face-to-face or remote meetings. We agreed to hold (remote) meetings of the subject areas organized by their coordinators every 2 months between the collaboration meetings. Coordinators meetings are organized on demand.
  • The first intermediate report is due in April. We will start to work on the general part in February. Thomas will clarify the scope.
  • The BMBF call for astro (particle) physics mentions research data management. How this relates to IDT-UM should be clarified.

Report for Area A by Manuel Giffels

  • The COBalD/TARDIS workshop was successful.
  • Groups are currently working separately on the topics. We should improve on synergy effects, e.g. with similar workshops.
  • It's difficult to find people for the positions.
  • Should columnar data analysis be included more?

Report for Area B by Christian Zeitnitz

  • Testing of components in Area B will probably come at a later stage in the project, currently focused on Area A.
  • Discussion about RUCIO: It's not an object store, works on the level of files. Should Germany increase its involvement in the project? Currently there is no funding, but it could be considered for future funding applications.

Report for Area C by Gregor Kasieczka

  • There are rumors about a change from Nvidia to Intel in the US. High level applications should be independent as long as low-level libraries adapt.
  • How to benefit from the collaboration, not only on paper? Can we generalize algorithms?
  • We should come up with a list of criteria/recommendations for different methods and establish a database of problems.

Report for Area D by Florian Bernlochner

  • The focus so far is on ACTS.

Bonn Application by Peter Wienemann

  • The collaboration expressed strong support for the proposal. The evaluation of techniques to reduce the site configuration and maintenance effort is addressing one of the main goals of the collaboration and is so far not covered for Tier 3 sites. The plan to disseminate the developed solutions is highly welcome as it adds an aspect that is essential for the success of the collaboration.
  • The focus is not on the development of new tools, but the evaluation of existing ones. It makes sense to start with the ones developed elsewhere in the collaboration.

Mainz Application by Andre Brinckmann

  • The collaboration sees the participation of a HPC center very positive and suggests to strengthen this unique feature in the application. While the technical aspects are certainly of interest, the main added value could come from the connection to the HPC world.
  • Several technical items were discussed showing the interest and the need for a closer communication. Interconnection is not a precondition, but useful for the proposed ad-hoc file system. Data stays for a scheduled time, but access is always possible so that it can be used as opportunistic resource for backfilling. Often jobs are I/O limited and thus need the performance the file system can provide.

ErUM-Data Digitization Board by Martin Erdmann

  • ErUM-Data is a chance for global, long-term funding for computing.

PAHN-PaN by Thomas Schoerner-Sadenius

  • NFDI has a slightly different emphasis, but it is not completely clear what will be funded by NFDI and what by ErUM-Data. We have to be pragmatic. We will know what is funded by NFDI when we have to write the next ErUM-Data proposal.

Computing/Software in Astrophysics by Matthias Steinmetz

  • Astrophysics is very advanced in open data. The main advantage to participate in a project is not anymore early access to data, but gaining know-how/experience.

Many thanks to Manuel and colleagues for the excellent local organization.

 

Es gibt eine zugehörige Notiz zu dieser Veranstaltung Anzeigen.