LAPACK Archives

[Lapack] ZGELS v.s. ZGESV

Hi.
  Thank you.

On Fri, 24 Aug 2007, Julien Langou wrote:
4- If the matrix is ill-conditioned ....
.......   They will filter the small singular values out of your 
initial matrix A, once those are removed your system becomes overdetermined 
and we solve a least squares problem on it. Those routines are ZGELSY (QR 
with column pivoting), ZGELSD (SVD with Divide and Conquer), ZGELSS (SVD with 
QR).

That what I was afraid ZGELS will silently do for me...
But now from your explanation I understood such an algorithm has a 
different driver routine .

Thanks, regards,
  Dima.

<Prev in Thread] Current Thread [Next in Thread>


For additional information you may use the LAPACK/ScaLAPACK Forum.
Or one of the mailing lists, or