Different LAPACK routines (and different algorithms) are used for real-symmetric and unsymmetric matrices. The former sorts the eigenvalues in ascending order, while the latter provides few guarantees about ordering.
On Friday, April 25, 2014 7:37:16 PM UTC-4, Ethan Anderes wrote: > > Is there a reason why the following two calls to eigs order the > eigenvalues differently? The first `mat1` is numerically not symmetric but > I would have still guessed the ordering of the magnitude of the eigenvalues > would be the same as those for `mat2`? > >
