Programming Abstractions for Run-Time Partitioning of Scientific Continuum Calculations Running on Multiprocessors

Programming Abstractions for Run-Time Partitioning of Scientific Continuum Calculations Running on Multiprocessors
复制标题

在多处理器上运行的科学连续体计算的运行时分区的编程抽象

DOI:
10.1007/3-540-45067-x_30
复制
发表时间:
1987
期刊:
26th Annual Symposium on Foundations of Computer Science (sfcs 1985)
影响因子:
--
通讯作者:
S. Baden
S. Baden
中科院分区:
--
文献类型:
--
作者:
S. Baden

文献摘要

被引文献

相似文献

。我将讨论在一组处理器上实现各种数学-物理计算的一组软件抽象。我尝试了Anderson局部修正方法的抽象,这是计算流体力学中的一种涡流方法。我在Intel iPSC(一种消息传递超立方体架构)的32个处理器和Cray X-MP(一种共享内存向量架构)的4个处理器上进行了实验,分别获得了24%和3.6%的良好并行加速比。抽象应该适用于不同的应用程序,包括有限差分方法,以及不同的体系结构,而不需要为每个新的体系结构对应用程序进行广泛的重新编程。
. I will discuss a set of software abstractions for implementing various math-physics cal culations on a team of processors. I tried out the abstractions on Anderson's Method of Local Corrections, a type of vortex method for computational fluid dynamics. I ran experiments on 32 processors of the Intel iPSC-a message-passing hypercube architecture-and on 4 processors of a Cray X-MP-a shared memory vector architecture - and achieved good parallel speedups of 24 and 3.6, respectively. The abstractions should apply to diverse applications, including finite difference methods, and to diverse archi tectures without requiring that the application be reprogrammed extensively for each new architecture.