iLCDirac and CI: Automated Testing for Distributed Computing

Abstract

Detector optimization studies for future high-energy physics experiments require the simulation and reconstruction of many physics processes and detector geometries. As an efficient way of accessing the necessary computational and storage resources, DIRAC has been developed and extended to form iLCDirac, which is specialized for the applications used in the context of linear collider detector studies. We give a short introduction of grid computing and the concept of high-throughput computing behind DIRAC before explaining the unique features of DIRAC and iLCDirac. With this preparation we explain how we leveraged continuous integration to provide smooth day-to-day operations and ensure that changes to the underlying code base do not cause an interruption of the service.