[petsc-users] Slepc: B not semi-positive definite

[Andrea Fani]

Hello,

I have to solve a  generalized eigenvalue problem Ax=sigma B x.

A is not hermitian. B is singular and is slightly not hermitian (only few
elements breaks the symmetry).

I used the standard  krilov-shur algorithm with a shift and invert
tranformation and it seems to work fine, but I am a bit worried since in
the documentation it is treated only the case with B singular but symmetric.

Andrea
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20140804/935bcfd7/attachment.html>