ZMCintegral:一个基于多GPU的多维Monte Carlo积分软件包
ZMCintegral: a Package for Multi-Dimensional Monte Carlo Integration on Multi-GPUs
摘要:
我们已经开发了一个Python包ZMCintegral,用于在多个图形处理单元(GPU)上进行多维Monte Carlo集成。该软件包采用分层抽样和启发式树搜索算法。我们已经构建了这个包的三个版本:一个使用Tensorflow,另两个使用Numba,两者都支持通用的用户定义函数,具有用户友好的界面。我们已经证明,Tensorflow和Numba可以帮助缺乏经验的科学研究人员在多个GPU上并行化他们的程序,而且工作量很小。对6维振荡函数和9维高斯函数这两种典型的被积函数进行了计算,并与VEGAS软件的计算精度和计算速度进行了比较。结果表明,ZMC积分的速度与VEGAS积分的速度相当,且具有一定的精度。对于繁重的计算,该算法可以在GPU的分布式集群上扩展。
英文摘要:
We have developed a Python package ZMCintegral for multi-dimensional Monte Carlo integration on multiple Graphics Processing Units(GPUs). The package employs a stratified sampling and heuristic tree search algorithm. We have built three versions of this package: one with Tensorflow and other two with Numba, and both support general user defined functions with a user-friendly interface. We have demonstrated that Tensorflow and Numba help inexperienced scientific researchers to parallelize their programs on multiple GPUs with little work. The precision and speed of our package is compared with that of VEGAS for two typical integrands, a 6-dimensional oscillating function and a 9-dimensional Gaussian function. The results show that the speed of ZMCintegral is comparable to that of the VEGAS with a given precision. For heavy calculations, the algorithm can be scaled on distributed clusters of GPUs.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。