Jul 16 – 26, 2022
US/Pacific timezone

Celeritas: : HEP detector simulation on GPUs

Jul 18, 2022, 7:00 PM
2h 20m
211 South Ballroom (HUB)

211 South Ballroom



Seth R. Johnson (Oak Ridge National Laboratory)


The previous Particle Physics Project Prioritization Panel (P5) report was responsible for consolidating a set of long term High Energy Physics (HEP) programs to address scientific questions on all three Department of Energy (DOE) Cosmic, Energy, and Intensity Frontiers. Two of these efforts are the High Luminosity Large Hadron Collider (HL-LHC) and its main experiments, and the Deep Underground Neutrino Experiment (DUNE). The unprecedented scale of these endeavors demands equally challenging computing capacity and storage requirements, with a commensurate fraction of the total computing cost being driven by Monte Carlo (MC) detector simulations. To alleviate this bottleneck, we present Celeritas, a new GPU MC detector simulation code designed to take advantage of the massive processing power of the DOE's Leadership Computing Facilities (LCFs). With Celeritas we plan to bridge the gap between HEP computing frameworks and the expanding DOE LCF network, vastly increasing the total compute capacity available to experiments for MC production campaigns. Here we present a roadmap for Celeritas, including its architecture, physics capabilities, and strategies for its integration with existing and future experimental HEP computing workflows.

In-person or Virtual? In-person

Primary authors

Amanda L. Lund (Argonne National Laboratory) Paul K. Romano (Argonne National Laboratory) Seth R. Johnson (Oak Ridge National Laboratory) Stefano C. Tognini (Oak Ridge National Laboratory) Guilherme Lima (Fermi National Accelerator Laboratory) Philippe Canal (Fermi National Accelerator Laboratory) Soon Yung Jun (Fermi National Accelerator Laboratory) Thomas M. Evans (Oak Ridge National Laboratory) Vincent R. Pascuzzi (Brookhaven National Laboratory)

Presentation materials