The EIC Software Stack: Designing a Scientific Software Environment for the 2030s
ORAL
Abstract
The Electron-Ion Collider aims to start data taking in the 2030s. The EIC will be the first accelerator designed and built during the era of artificial intelligence, machine learning, streaming readout, and ubiquitous access to GPU and FPGA code accelerators at distributed computing clusters worldwide. The EIC Detector Collaboration has engaged in a deliberative user-centered design process to develop the nuclear physics software stack of the future, built for continuity, performance, flexibility and sustainability. Based on the EIC Statement of Software Principles, we have chosen several best-in-class software components from geometry description (DD4hep), data description (podio/EDM4hep), and event reconstruction, through code repository (GitHub) and continuous integration. In order to prepare the workforce for these technologies, we have developed a training plan for new researchers joining the EIC efforts or outside initiatives aiming to take advantage of the ecosystem we are developing.
–
Presenters
-
Wouter Deconinck
Univ of Manitoba
Authors
-
Wouter Deconinck
Univ of Manitoba
-
Sylvester J Joosten
Argonne National Laboratory
-
Markus Diefenthaler
Jefferson Lab/Jefferson Science Associat
-
Zhoudunming Tu
Brookhaven National Laboratory
-
Wenliang B Li
Stony Brook University, Center for Frontiers in Nuclear Science, Stony Brook University
-
Cristiano Fanelli
Massachusetts Institute of Technology MIT
-
David J Lawrence
Jefferson Lab
-
Joe Osborn
Oak Ridge National Lab
-
Andrea Bressan
INFN Trieste
-
Tanja Horn
Catholic Univ of America
-
Torre J Wenaus
Brookhaven National Laboratory
-
Kolja Kauder
Brookhaven National Laboratory
-
Whitney R Armstrong
Argonne National Laboratory
-
Cameron Dean
Brookhaven National Laboratory