Abstract—A widely used Monte Carlo event generator is A Multi-Phase Transport model (AMPT) for relativistic heavy-ion collisions. It depends on Zhang’s Parton Cascade (ZPC) package to simulate initial stage parton cascade. Based on ZPC, we have developed a code for the simulation of the parton cascade to exploit the powerful parallel processing capability of GPU. The goal is to accelerate the simulation of the parton cascade in a system of partons that is formed in ultrarelativistic heavy-ion collisions. Named PCG (Parton Cascade on GPU), the code makes real time collision detection among N interacting partons formed in a heavy-ion collision parallelized. The parallelization was implemented by using CUDA C. With simulating Pb-Pb collisions at sqrt(sNN)=2.76 TeV as a use case, we first verified the correctness of PCG through comparison of the output of PCG with those of ZPC, then we estimated the computational efficiency of PCG to be 2x to 3x relative to ZPC, which is a serial code and only runs on CPU. Therefore PCG is viable for being integrating into AMPT for simulating heavy-ion collisions and can save large amount of computing resources for large scale AMPT-based event generation in ultrarelativistic heavy-ion collisions at sqrt(sNN)=2.76 TeV.
Index Terms—GPU, CUDA C, simulation of parton cascade, ultrarelativistic heavy-ion collision.
Qingjun Liu is with the Beijing Institute of Petro-chemical Technology, Beijing 102617 China (e-mail: liuqingjun@bipt.edu.cn).
Weiqin Zhao is with the Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 China (e-mail: zhaowq@ihep.ac.cn).
Fang Liu, Ningming Nie, and Chunbao Zhou are with the Supercomputer Center, Computer Network Information Center, Chinese Academy of Sciences, Beijing 100190 China (e-mail: liuf@sccas.cn, nienm@sccas.cn, zhoucb@sccas.cn).
[PDF]
Cite: Qingjun Liu, Weiqin Zhao, Fang Liu, Ningming Nie, and Chunbao Zhou, "GPU-Accelerated Parton Cascade in Heavy-Ion Collisions," International Journal of Computer Theory and Engineering vol. 8, no. 6, pp. 439-443, 2016.