Scalable and massively parallel Monte Carlo photon transport simulations for heterogeneous computing platforms
Department of ECE, Northeastern University, 360 Huntington Ave, Boston, MA, 02115, U.S.A.
arXiv:1711.03244 [cs.DC], (9 Nov 2017)
@article{yu2017scalable,
title={Scalable and massively parallel Monte Carlo photon transport simulations for heterogeneous computing platforms},
author={Yu, Leiming and Nina-Paravecino, Fanny and Kaeli, David and Fang, Qianqian},
year={2017},
month={nov},
archivePrefix={"arXiv"},
primaryClass={cs.DC}
}
We present a highly scalable Monte Carlo (MC) 3D photon transport simulation platform designed for heterogeneous computing systems. By developing a massively parallel MC algorithm using the OpenCL framework, this research extends our existing GPU-accelerated MC technique to a highly-scalable vendor-independent heterogeneous computing environment, achieving significantly improved performance and software portability. A number of parallel computing techniques are investigated to achieve portable performance over a wide range of computing hardware. Furthermore, multiple thread-level and device-level load-balancing strategies have been developed to obtain efficient simulations using multiple CPUs and GPUs.
November 12, 2017 by hgpu
