VecGeom: A new vectorised geometry library for particle-detector simulation

Not scheduled
15m
OIST

OIST

1919-1 Tancha, Onna-son, Kunigami-gun Okinawa, Japan 904-0495
poster presentation Track8: Performance increase and optimization exploiting hardware features

Speaker

Sandro Christian Wenzel (CERN)

Description

A geometry modeller library is among the most important components of the software simulating the passage of particles in a detector, and many experiment simulations are currently based on the geometry implementations offered by Geant4 or ROOT. Here, we report on our effort to extend, re-engineer and evolve thes libraries in multiple directions in order to make them ready for the future challenges of computing in HEP. This includes primarily an extended API as well as SIMD-vectorised algorithms able to efficiently handle geometry queries for multiple particles at the same time. Secondly, we aim for a native support of the geometry module on the GPU or on mixed heterogeneous CPU/GPU platforms. This effort is one of the essential ingredients of the Geant-Vector project that is focusing on a fine-grained multithreaded and vectorised design, propagating many particles from different events at the same time. Our presentation will give an overview of the new vectorizsed, generic and templated geometry library "VecGeom" that accomplishes those primary goals and which also improves the overall geometry performance for current simulation frameworks. Beyond discussing new performance numbers for elementary and higher level geometry algorithms, we will show a first global evaluation of the new library on a realistic detector in form of the geometry description from the CMS experiment

Primary author

Co-authors

Andrei Gheata (CERN) Mr Federico Carminati (CERN) Dr Gabriele Cosmo (CERN) Georgios Bitzes (National and Kapodistrian University of Athens (GR)) Guilherme Lima (FermiLab (US)) Johannes Christof de Fine Licht (University of Copenhagen (DK)) John Apostolakis (CERN) Laurent Duhem Marilena Bandieramonte (Universita e INFN (IT)) Mihaly Novak (CERN) Oksana Shadura (National Technical Univ. of Ukraine "Kyiv Polytechnic Institute) Philippe Canal (Fermi National Accelerator Lab. (US)) Raman Sehgal (Bhabha Atomic Research Centre (IN)) Dr Rene Brun (CERN) Soon Yung Jun (Fermi National Accelerator Lab. (US)) Tatiana Nikitina (Gangneung-Wonju National University (KR)) Victor Daniel Elvira (Fermi National Accelerator Lab. (US))

Presentation materials