[Lapack] Extended precision arithmetic with LAPACK
From: Dimitris Giannakis
Date: Tue, 27 Feb 2007 10:29:49 -0600
Apologies if this question has been asked before, but I could not
find a relevant discussion in the forum.
I am a Physics PhD student and, as part of hydrodynamic stability
calculations, I need to solve generalized eigenvalue problems A x =
\lambda B x, where the matrix A is highly non-normal. It has been
documented in the literature (e.g. Dongarra et al, Appl. Numer. Math.
22, 399 (1996)) that 128 bit arithmetic is often required to obtain
numerically accurate results for this type of problems.
I was therefore hoping to compile LAPACK to work with 128 bit
arithmetic (or any other format that is more precise than double),
and would be most appreciative if you could point me to any
guidelines/information available on how to do this..
My options are to either work on my home computer (an Apple G4
running OS X), or to use an account on a Linux machine with Intel
Xeon processors (I understand the latter option would be far more
efficient for extended precision arithmetic).
Thank you in advance for your assistance.
<Prev in Thread]
[Next in Thread>
[Lapack] Extended precision arithmetic with LAPACK,