symmetric tridiagonal eigenproblem

From: Gert Van den Eynde <gvdeynde@xxxxxxxxx>
Date: Tue, 01 Aug 2006 16:07:50 +0200

Hi all,

I have to solve two related symmetric tridiagonal eigenproblems. One of
size N, the other of size N+1 where the upper-left NxN submatrix is
identical to the former one (i.e. I add only one row and one column). Is
there a way to re-use the information from the N-size calculation to
speed up the N+1-size calculation? Preferably using LAPACK (I already
use DSTEVD for both problems separately).

Thanks,
gert
.

Follow-Ups:
- Re: symmetric tridiagonal eigenproblem
  - From: Rich Townsend
- Re: symmetric tridiagonal eigenproblem
  - From: Arnold Neumaier

Prev by Date: Re: Sin & Cos speed worries
Next by Date: Re: symmetric tridiagonal eigenproblem
Previous by thread: Problems with Franke's inverse distance weighting interpolation
Next by thread: Re: symmetric tridiagonal eigenproblem
Index(es):
- Date
- Thread

Relevant Pages

Re: symmetric tridiagonal eigenproblem
... Gert Van den Eynde wrote: ... I have to solve two related symmetric tridiagonal eigenproblems. ... the other of size N+1 where the upper-left NxN submatrix is ... there a way to re-use the information from the N-size calculation to ...
(sci.math.num-analysis)
Re: symmetric tridiagonal eigenproblem
... Gert Van den Eynde wrote: ... the other of size N+1 where the upper-left NxN submatrix is ... there a way to re-use the information from the N-size calculation to ... so you can use the method for rank 1 updates of eigenvalue problems ...
(sci.math.num-analysis)