Re: Eigensystem Bug? OS-X

*To*: mathgroup at smc.vnet.net*Subject*: [mg56138] Re: [mg56101] Eigensystem Bug? OS-X*From*: Sseziwa Mukasa <mukasa at jeol.com>*Date*: Sat, 16 Apr 2005 03:52:47 -0400 (EDT)*References*: <200504150847.EAA11575@smc.vnet.net>*Sender*: owner-wri-mathgroup at wolfram.com

On Apr 15, 2005, at 4:47 AM, Erich Mueller wrote: > I have some strange behavior of Eigensystem and Eigenvectors [but not > Eigenvalues] on my G5 running OS-X 10.3.8, and Mathematica 5.1.0.0. > > The Kernel Hangs when I try to calculate the Eigenvectors of the 18x18 > Matrix at the end of this message. [Its the smallest matrix for which I > have the problem.] It does not respond to an "Abort Evaluation", or > "Interupt Evaluation Signal", and I must kill it from the terminal. > > There is no problem finding the eigenvalues of this same matrix on a > PC. > Is there someone who can suggest a fix/workaround? Failing that, is > there anyone with a mac who can verify the problem? > I have the same problem in version 5 on Mac OS X 10.3.8. It is not a case of the front end taking a long time to format output because suppressing the output with ; still takes forever to compute. On the other hand the matrix is very ill conditioned, so reduction to Schur form using SchurDecomposition would probably be preferable. You can then find eigenvectors by inverse iteration. Regards, Ssezi

**References**:**Eigensystem Bug? OS-X***From:*Erich Mueller <em256@NOSPAM.cornell.edu>