【摘要】 分析了不同并行伪随机数生成策略的优劣,结合线性同余(linear congruence generator LCG)算法内在属性及机群系统主从(m/w)计算模型,提出并行伪随机数生成器DLLCG(double level LCG)算法。利用具体Monte Carlo仿真应用,验证了该并行随机数生成算法在各类机群系统上的可用性。通过理论和实验分析,表明该算法可以用于同构或异构机群系统中,具有良好的可扩展性和移植性。更多还原

【Abstract】 The strategies for parallel pseudo-random number generator were discussed. Combining the property of linear congruence generator(LCG) with cluster master/ worker(m/w) computing model,a parallel pseudo-random number generator,DLLCG(double level LCG) was designed. From theory analyzing and practical Monte Carlo simulation application,the DLLCG’s quality was proved. It can be used in homogeneous and heterogeneous cluster.更多还原