Re: ? open SVD f90 code

From: Rich Townsend (rhdt_at_barVOIDtol.udel.edu)
Date: 10/21/04 

Date: Thu, 21 Oct 2004 11:31:14 -0400

Jugoslav Dujic wrote:
> Cheng Cosine wrote:
> | "Jugoslav Dujic" <jdujic@yahoo.com> wrote in message
> | news:2tkoaaF2291mmU1@uni-berlin.de...
> || Cheng Cosine wrote:
> ||| Though Fortran Recipes offers SVD code but that only works for real var.
> |||
> ||| I tried to d/l codes from netlib but my compiler, LF90v45, does not
> |||
> ||| compile it, but only gives me a series of errors. The package from
> |||
> ||| netlib contains many files so I dare not to trace them.
> |||
> ||| What are other site(s) I can d/l SVD alg that deals both real and complex
> |||
> ||| variables?
> ||
> || Which "netlib package"?
> ||
> || I'm not too familiar with SVD, but I did download some related LAPACK
> || routines (www.netlib.org/lapack) which compiled & worked just fine. Go to
> ||
> || http://www.netlib.org/lapack/individualroutines.html
> ||
> || choose your datatype, then search for "svd" there.
> ||
> |
> | I guess we have the same files. One of the file is called
> |
> | zgesvd.f90
> |
> | I guess it's also from LAPACK you mentioned. But LF90v45 does not like it.
> | Neither does the free compiler in our school's computer center.
>
> Hmmm. LAPACK is one of the most nicely & most standardly written codes I've
> seen.
> Could it possibly be that Lahey doesn't support complex*16 (or at least
> that particular syntax)?
>

I've had no problems compiling LAPACK using the LF95 compiler (v 6.2c
for Linux). The LF90 product is pretty old.

Its worth mentioning that LAPACK is pure Fortran 77. There is a Fortran
95 wrapper to LAPACK, LAPACK95, but all of the routines contained in it
have the prefix la_. Therefore, I don't know where the 'zgesvd.f90'
routine is coming from -- but it doesn't belong to either library.

cheers,

Rich 

-- 
Dr Richard H D Townsend
Bartol Research Institute
University of Delaware
[ Delete VOID for valid email address ]

Relevant Pages

  • Re: Who uses clapack?
    ... I have used Lapack as a reference to write a Divide and Conquer ... or C++ which are just as suitable as Fortran. ... No dynamic memory allocation but the eternal WORK and IWORK ... Lapack's use of a BLAS is great but Lapack is tailored to the ...
    (comp.lang.fortran)
  • Re: LAPACK
    ... >>what do I have to do for installing LAPACK on a pc with windows XP??? ... You need a Fortran compiler, unless you want to use CLAPACK. ... The speed of LAPACK is very dependent on the quality of your BLAS. ...
    (sci.math.num-analysis)
  • Re: Who uses clapack?
    ... I have used Lapack as a reference to write a Divide and Conquer ... or C++ which are just as suitable as Fortran. ... No dynamic memory allocation but the eternal WORK and IWORK ... Lapack's use of a BLAS is great but Lapack is tailored to the ...
    (sci.math.num-analysis)
  • Re: Singular Value Decomposition (non-square matrices)?
    ... Value Decomposition (SVD) of non-square matrices? ... There is an Ada binding for BLAS, ... There is apparently a partial Ada binding to LAPACK; ...
    (comp.lang.ada)
  • Re: Lapack/ CLapack
    ... > I need to convert a Fortran program I wrote some months ago. ... is it better to use the original LAPACK compiled from Fortran ... > less performant? ... probably just one for BLAS and one for LAPACK. ...
    (comp.lang.c)