Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!ucsd!ucbvax!durham.ac.uk!Lydia.Heck From: Lydia.Heck@durham.ac.uk Newsgroups: comp.sys.transputer Subject: parallelised code for matrix-manipulations Message-ID: Date: 20 Aug 90 21:10:14 GMT Sender: daemon@ucbvax.BERKELEY.EDU Organization: The Internet Lines: 16 Our group in Durham (UK) would need a parallelised code for the general eigenvalue problem A.X = lambda B. X where A and B are matrices and lambda a scalar using the Cholesky factorisation method. We run our codes in fortran on a Meiko computing surface using an occam-harness (Fortnet) Is there anybody out there who knows about parallelised code for such a problem, then please contact me, either via the mailing list or directly via the e-mail address Lydia.Heck@uk.ac.durham Thank you P.s.: We run the Liverpool library, however to find all eigenvalues for a matrix using the appropriate program of that library is cumber- some as one has to guess each eigenvalue in advance; this gets a bit tedious for any sensible problem we are considering.