[petsc-users] Solving tridiagonal hermitian generalized eigenvalue problem with SLEPC

[Jed Brown]

Hong <hzhang at mcs.anl.gov> writes:

> Toon :
>> Sorry, I forgot to mention: I am looking for the Eigenvalues that are the
>> largest, not in absolute value, but along the real axis.
>
> Then, you do not need shift-invert, therefore, should not use LU
> matrix factorization.

The error is in STSetUp_Shift, not STSetUp_Sinvert.  This is a
generalized eigenvalue problem.  Of course you can change the PC, but I
think we should understand the problem first.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 818 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20140818/04bccadb/attachment.pgp>