Title :
Parallel methods for iterative matrix decompositions
Author_Institution :
Dept. of Comput. Sci., Yale Univ., New Haven, CT, USA
Abstract :
Parallel methods for iterative matrix decompositions are developed by using the possibility of approximate (but still orthogonal) rotations together with suitable kinds of rotation schemes. Favorable choices of some of these approximations are applied together with suitable rotation schemes (factorized rotation scheme, shift-and-add rotation scheme) in order to obtain algorithms highly suited for parallel implementations. The author mainly deals with the Jacobi method for the symmetric eigenvalue problem, but the ideas presented can also be applied to other iterative matrix decomposition algorithms using orthogonal rotations
Keywords :
convergence of numerical methods; eigenvalues and eigenfunctions; iterative methods; mathematics computing; matrix algebra; parallel algorithms; Jacobi method; decomposition algorithms; factorized rotation scheme; iterative matrix decompositions; orthogonal rotations; parallel implementations; symmetric eigenvalue problem; Computer science; Concurrent computing; Convergence; Eigenvalues and eigenfunctions; Iterative algorithms; Iterative methods; Jacobian matrices; Matrix decomposition; Process design; Symmetric matrices;
Conference_Titel :
Circuits and Systems, 1991., IEEE International Sympoisum on
Print_ISBN :
0-7803-0050-5
DOI :
10.1109/ISCAS.1991.176316
