Verification of Electromagnetic Physics Models for Parallel Computing Architectures in the GeantV Project

G Amádio ,J Apostolakis ,F Carminati ,R Seghal ,L Duhem,Mihály Novák ,Shibayama Jun ,D Elvira ,Ivantchenko ,S P Behera ,S Vallecorsa ,A Ribon ,R Brun ,T Nikitina ,Sandro Christian Wenzel ,Philippe Canal ,I Goulas ,D Konstantinov ,Witold Pokorski ,A Gheaţă ,Mihaela Gheata ,G Cosmo ,Farah Hariri ,G Folger ,Marilena Bandieramonte ,H Kumawat ,O Shadura ,J G Rocha De Lima

doi:10.1088/1742-6596/898/4/042019

Abstract

An intensive R&D and programming effort is required to accomplish new challenges posed by future experimental high-energy particle physics (HEP) programs. The GeantV project aims to narrow the gap between the performance of the existing HEP detector simulation software and the ideal performance achievable, exploiting latest advances in computing technology. The project has developed a particle detector simulation prototype capable of transporting in parallel particles in complex geometries exploiting instruction level microparallelism (SIMD and SIMT), task-level parallelism (multithreading) and high-level parallelism (MPI), leveraging both the multi-core and the many-core opportunities. We present preliminary verification results concerning the electromagnetic (EM) physics models developed for parallel computing architectures within the GeantV project. In order to exploit the potential of vectorization and accelerators and to make the physics model effectively parallelizable, advanced sampling techniques have been implemented and tested. In this paper we introduce a set of automated statistical tests in order to verify the vectorized models by checking their consistency with the corresponding Geant4 models and to validate them against experimental data.

Highlights

High-energy particle physics has advanced greatly over recent years and current plans for the future foresee even more ambitious targets and challenges that have to be coped with
We present preliminary verification results concerning the electromagnetic (EM) physics models developed for parallel computing architectures within the GeantV project, focusing in particular, as a case study, on the verification of the KleinNishina model for Compton process implemented in GeantV, against the corresponding model extracted from the Geant4 implementation
This makes it necessary to develop a Vectorized Physics library (VecPhys) which, following the parallel approach to the transport of particles proposed in GeantV, is capable of processing several tracks at a time to profit from the instruction level parallelism

Summary

Introduction

High-energy particle physics has advanced greatly over recent years and current plans for the future foresee even more ambitious targets and challenges that have to be coped with. A physics library for parallel architectures In the context of a typical HEP simulation, a considerable fraction (30%-50%) of the time is spent in executing algorithms that sample cross sections and implement the physical processes This makes it necessary to develop a Vectorized Physics library (VecPhys) which, following the parallel approach to the transport of particles proposed in GeantV, is capable of processing several tracks at a time to profit from the instruction level parallelism. For this purpose we have been developing a statistical verification and validation suite to compare, test and validate all the relevant physical quantities of every specific physics process It consists of different automated regression analysis tests that can be run on the results of a simulation to identify deficiencies of algorithms or errors in implementation. The points mutual distribution provides additional statistical information on the form, outliers, skewness, kurtosis and other features of the distributions

Test case

Findings

Conclusions