首页> 外文会议>Proceedings of the 2007 International Conference on Parallel and Distributed Processing Techniques and Applications(PDPTA2007) >DistMe: A Java Based Toolbox for the Distribution of Parallel Monte Carlo Simulations. Application to Nuclear Medicine Using the GATE Simulation Package
【24h】

DistMe: A Java Based Toolbox for the Distribution of Parallel Monte Carlo Simulations. Application to Nuclear Medicine Using the GATE Simulation Package

机译:DistMe:用于分发并行Monte Carlo模拟的基于Java的工具箱。使用GATE仿真包在核医学中的应用

获取原文

摘要

Monte Carlo simulations needing many replicates to obtain good statistical results can be executed in parallel using the "Multiple Replications In Parallel" approach using parallel streams of pseudo-random numbers. We have built a generic object-oriented toolbox called DistMe to automate such parallelization for any Monte Carlo simulation. This toolbox, written in Java is freely available on SourceForge, it is based on the use of a documented XML format for random numbers generators statuses. We have used this tool in a concrete case of distributed tomographic reconstruction using local clusters and grid computing. We obtained very convincing results with this large medical application, achieving in one week computations that could have taken more than 3 years of processing on a single computer.
机译:可以使用伪随机数的并行流,使用“并行多次复制”方法并行执行需要大量重复才能获得良好统计结果的蒙特卡洛模拟。我们已经构建了一个称为DistMe的面向对象的通用工具箱,以针对任何Monte Carlo模拟自动进行这种并行化。该工具箱用Java编写,可在SourceForge上免费获得,它基于已记录的XML格式用于随机数生成器状态。我们在使用局部聚类和网格计算的分布式断层摄影重建的具体情况下使用了该工具。在这种大型医疗应用中,我们获得了令人信服的结果,在一星期内完成的计算可能在一台计算机上花费了3年以上的处理时间。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号