基于DIRAC的高能物理分布式计算系统在CEPC实验上的应用

The Distributed Computing System Based on DIRAC for High Energy Physics and Its Application in CEPC Experiment

2 0 1 2年L H C实验发现希格斯粒子后,我国粒子物理学界规划了高能环形正负电子对撞机(CEPC),以及它升级后的超级质子对撞机(Spp C),用于研究希格斯物理,并在高能量前沿探索超越标准模型的新物理。CEPC实验的计算量非常大,即使是在预研阶段,一次探测器全模拟就需要833万CPU小时的计算量和1PB存储空间。当本地计算资源不足时,分布式计算是一种快速整合合作单位计算资源来满足实验大量计算需求的有效方法。2012年,高能所基于DIRAC中间件建立了一套约2000 CPU核规模的分布式计算系统,服务于BESIII实验。本文介绍了将此分布式计算系统扩展到支持CEPC实验的改进方案,并介绍了此系统在CEPC实验的实际应用。

After the discovery of Higgs bosons in 2012 at LHC, China＇s particle physics community planed a high energy Circular Electron Positron Collider （CEPC）, together with its upgrade form Super positron-positron Collider （SppC）, to investigate Higgs physics and search new physics beyond the standard model. CEPC experiment need plenty of computing resources, even in pre-study stage,a full simulation of detector requires 350 000 CPU-days computing and 1PB storage. In case the local resource cannot fulfill the experiment＇s requirements, the distributed computing is a way of fast collecting the resources of collaboration members and serve the experiment＇s computing. In Institute of High Energy Physics, a distributed computing system based on DIRAC middleware was built for BESIII experiment in 2012. In this paper, we present an improvement scheme of extending this distributed computing system to support CEPC experiment, together with its implementation and application.