Heterogeneous computing is becoming increasingly important for efficient scientific computing. Portable and scalable computing paradigms are in demand to leverage acceleration hardware at distributed data centers and High-Performance Computing centers. This is a one-day in-person and virtual (zoom) mini-workshop to discuss the heterogeneous computing ‘as-a-service’ solution. The workshop starts with an overview of heterogeneous computing ‘as-a-service’ enabled by SONIC in particle/astroparticle physics, and astrophysics experiments such as ATLAS and CMS, DUNE, LIGO, and IceCube. We also aim to discuss 'inference-as-a-service' server support development with industry partners such as NVIDIA Triton, Google TPU, and AMD.
Organizing Committee:
Martin Kwok (U. Nebraska-Lincoln)
Mia Liu (Purdue)
Phil Harris (MIT)
Kevin Pedro (Fermilab)
Giuseppe Cerati (Fermilab)
Michael Wang (Fermilab)
Benedikt Riedel (U. Wisconsin-Madison)
Daniel Diaz (UCSD)
LPC Events Committee:
Gabriele Benelli (Brown, Co-chair)
Abhijith Gandrakota (Fermilab, Co-chair)
LPC Coordinators:
James Hirschauer (Fermilab)
Isobel Ojalvo (Princeton)
This event is sponsored by the Accelerated AI Algorithms for Data-Driven Discovery (A3D3) Institute and the LPC Physics Center (LPC)
