摘要
三维连续截面多粒子输运蒙特卡罗程序MCNP-4C-经过MPI并行改造,实现了MPI 并行化.采用分段随机数发生器,并行取得了与串行完全一致的结果,500个处理器的计算速度较串行提高了460倍,并行效率达到92%,可计算包括临界在内的多粒子输运问题.
The parallelization of MCNP-4C Monte Carlo N-Particle Transport code system in MPI has been realized by modifying the serial code. The parallel results are the same with the serial results by using the segment random number generator. The parallel speed of 500-processors is 460-times faster than the serial speed. The parallel efficiency is up to 92%. The parallel code can simulate multi-particle transport problems that include the critical problems.
出处
《数值计算与计算机应用》
CSCD
2006年第1期52-59,共8页
Journal on Numerical Methods and Computer Applications
基金
计算物理实验室基金中国工程物理研究院基金NSAF联合基金(10576006)资助
关键词
粒子输运
蒙特卡罗
MPI并行化
分段随机数
particle transport, Monte Carlo, parallelization, MPI, segment random number