Iterative methods for solving linear systems of equations on FPGA-based machines

被引:0
|
作者
Xu, XZ [1 ]
Ziavras, SG [1 ]
机构
[1] New Jersey Inst Technol, ECE Dept, Newark, NJ 07102 USA
关键词
FPGA; iterative methods; Newton-Raphson; conjugate gradient; Chebyshev iteration;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
This paper focuses on the solution to a system of linear equations by iterative methods. A short survey is presented on direct and iterative methods. Three practical iterative methods are discussed. Two preconditioning techniques are employed to speed up the iterations. Two testbeds are used to carry out a comparison of these iterative methods. For the Del18200 PC testbed, we show the pros and cons of the iterative methods; for an Altera(TM) NIOS-based reconfigurable computing system with FPU extensions, we prove the efficiency of our FPU extensions, which can significantly speedup the iterative calculations.
引用
收藏
页码:472 / 475
页数:4
相关论文
共 50 条