...
首页> 外文期刊>Computer physics communications >3D ELECTROMAGNETIC PLASMA PARTICLE SIMULATIONS ON A MIMD PARALLEL COMPUTER
【24h】

3D ELECTROMAGNETIC PLASMA PARTICLE SIMULATIONS ON A MIMD PARALLEL COMPUTER

机译:MIMD并行计算机上的3D电磁等离子体粒子模拟

获取原文
获取原文并翻译 | 示例
           

摘要

A three-dimensional electromagnetic PIC code has been developed on the 512 node Intel Touchstone Delta MIMD parallel computer. This code uses a standard relativistic leapfrog scheme to push particles and a local finite-difference time-domain method to update the electromagnetic fields. The code is implemented using the General Concurrent PIC algorithm which uses a domain decomposition to divide the computation among the processors. The 3D simulation domain can be partitioned into 1-, 2-, or 3-dimensional subdomains. Particles must be exchanged between processors as they move among the subdomains, The Intel Delta allows one to use this code for very-large-scale simulations (i.e. over 10(8) particles and 10(6) grid cells). The parallel efficiency of this code is measured, and the overall code performance on the Delta is compared with that on Gray supercomputers. It is shown that our code runs with a high parallel efficiency of greater than or equal to 95% for large size problems. The particle push time achieved is 115 ns/particle/time step for 162 million particles on 512 nodes. Compared with the performance on a single CPU Gray C90, this represents a factor of 58 speedup. It is also shown that the finite-difference method for the field solve is significantly more efficient than transform methods on parallel computers. The field solve time is < 0.7% of total time for problems with 77 particles/cell, and it is < 3% even for problems with 7 particles/cell. [References: 16]
机译:在512节点Intel Touchstone Delta MIMD并行计算机上开发了三维电磁PIC代码。该代码使用标准的相对论越级方案来推动粒子,并使用局部时差有限域法来更新电磁场。该代码使用通用并发PIC算法实现,该算法使用域分解在处理器之间划分计算量。 3D模拟域可以划分为1、2或3维子域。当粒子在子域之间移动时,必须在处理器之间交换粒子。IntelDelta允许人们将此代码用于超大规模仿真(即超过10(8)个粒子和10(6)网格单元)。测量该代码的并行效率,并将Delta上的总体代码性能与Gray超级计算机上的总体性能进行比较。结果表明,对于大型问题,我们的代码以大于或等于95%的高并行效率运行。对于512个节点上的1.62亿个粒子,实现的粒子推动时间为115 ns /粒子/时间步长。与单颗CPU Gray C90的性能相比,它的速度提高了58倍。还表明,用于现场求解的有限差分方法比并行计算机上的变换方法有效得多。对于77个粒子/电池的问题,现场求解时间小于总时间的0.7%,对于7个粒子/电池的问题,其求解时间也小于3%。 [参考:16]

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