This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: LAPACK vs. GSL matrix algebra
- From: Brian Gough <bjg at network-theory dot co dot uk>
- To: gsl-discuss at sources dot redhat dot com
- Date: Thu, 30 Sep 2004 19:27:35 +0100
- Subject: Re: LAPACK vs. GSL matrix algebra
- References: <Pine.GSO.3.96.1040927123018.26477A-100000@yatsen.cas.McMaster.CA><1096308362.2446.67.camel@bellerophon.lanl.gov><20040927200259.GN22210@austin.ibm.com><1096318490.2446.107.camel@bellerophon.lanl.gov><20040929175417.GC11461@austin.ibm.com><1096494608.2446.186.camel@bellerophon.lanl.gov>
As a test case, see if you can wrap DGESVX (AX=B) in a meaningful way.
I think it always ends up as "XA=B" in C, regardless of TRANS='N' or
'T'.
It looks like Clint Whaley made his own implementation of the small
number of LAPACK routines in ATLAS so that they could have an
additional CBLAS_ORDER argument.
--
Brian Gough
Network Theory Ltd,
Commercial support for GSL --- http://www.network-theory.co.uk/gsl/