基于Jacobi算法对称矩阵特征值计算的FPGA实现  

Implementation in FPGAs of eigenvalue calculation with Jacobi algorithm for symmetric matrix

作　　者：袁生光[1] 沈海斌[1] 

机构地区：[1]浙江大学超大规模集成电路设计研究所,浙江杭州310027

出　　处：《机电工程》2008年第10期80-82,106,共4页Journal of Mechanical & Electrical Engineering

摘　　要：对称矩阵的特征值计算问题在数学和工程应用中都具有重要的实际意义,针对对称矩阵特征值计算这一问题提出了一种新的FPGA实现方案,并对该方案进行了验证和性能评估,最后,将它和目前已有的设计进行了速度与面积方面的对比。研究结果证实,该方案在速度不受影响的前提下,所占面积得到了减小。Eigenvalue calculation for symmetric matrix is significant in mathematics and engineering. A novel architecture of FPGAs was demonstrated to solve eigenvalue calculation for symmetric matrix. Moreover, the performance of the architecture was validated and evaluated, and finally it was compared with others mentioned in some articles. The experimental results prove that the area of the implementation is less than others without reducing speed.

关 键 词：对称矩阵 特征值计算 现场可编程门阵列 Jacobi算法 坐标施转数字计算机 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]