Title :
Hardware design of an eigensolver based on the QR method
Author :
Guerrero-Ramirez, J.E. ; Velasco-Medina, J. ; Arce-Clavijo, J.C.
Author_Institution :
Bionanoelectronics Res. Group, Univ. del Valle, Cali, Colombia
fDate :
Feb. 27 2013-March 1 2013
Abstract :
Eigenvalue computation is essential in many fields of science and engineering. In most cases we need a lot of computation resources in order to compute all the eigenvalues of medium and large matrices. The QR method is one of the most famous methods that computes all the eigenvalues and occasionally, the eigenvectors. In this paper we have designed a systolic architecture for the QR eigensolver, which implements the basic QR algorithm for symmetric matrices.
Keywords :
digital arithmetic; eigenvalues and eigenfunctions; matrix algebra; QR eigensolver; QR method; eigenvalue computation; eigenvectors; hardware design; symmetric matrices; systolic architecture; Algorithm design and analysis; Arrays; Eigenvalues and eigenfunctions; Hardware; Matrix decomposition; Symmetric matrices; Vectors; CORDIC; Eigenvalue computation; Givens rotations; QR method; Systolic Array;
Conference_Titel :
Circuits and Systems (LASCAS), 2013 IEEE Fourth Latin American Symposium on
Conference_Location :
Cusco
Print_ISBN :
978-1-4673-4897-3
DOI :
10.1109/LASCAS.2013.6519065
