[Coin-discuss] COIN-OR licences again...

Andreas Waechter andreasw at watson.ibm.com
Tue Apr 8 12:26:01 EDT 2008


> On Mon, 2008-04-07 at 08:00 -0600, Brian Borchers wrote:
>>> I don't see a problem in the lapack license... And even if there was
>>> one, atlas is the preferred (potentially faster) replacement.
>>
>> ATLAS implements the BLAS routines and a very small subset of the
>> LAPACK routines.  The vast majority of LAPACK routines aren't included
>> in ATLAS.
>
> Hmmhh, as I don't have to have lapack installed (linking to atlas libs
> is enough) to get all of the lapack routines, atlas must be simply
> sucking in the lapack routines for which it does not implement.

http://math-atlas.sourceforge.net/errata.html#completelp :

-------------- 8< -------------

Building a complete LAPACK library

ATLAS does not provide a full LAPACK library. However, there is a simple 
way to get ATLAS to provide its faster LAPACK routines to a full LAPACK 
library provided by netlib. First, install lapack from netlib. First, 
download and install the standard LAPACK library from the LAPACK homepage. 
Then when doing the configure step to ATLAS pass the following flag:

    --with-netlib-lapack=[path/to/lapack/lib]

If you need to add full LAPACK after ATLAS has already been built, in your 
BLDdir/lib/ directory (where you should have a liblapack.a), issue the 
following commands:

   mkdir tmp
   cd tmp
   ar x ../liblapack.a
   cp <your LAPACK path & lib> ../liblapack.a
   ar r ../liblapack.a *.o
   cd ..
   rm -rf tmp

Just linking in ATLAS's liblapack.a first will not get you the best LAPACK 
performance, mainly because LAPACK's untuned ILAENV will be used instead 
of ATLAS's tuned one. So, if you use any LAPACK routine that is not 
provided by ATLAS, it is essential that you create this hybrid 
LAPACK/ATLAS library in order to get the best performance.

-------------- 8< -------------

>
> Soeren
> -- 
> Soeren Sonnenburg - Fraunhofer FIRST      Tel: +49 (30) 6392 1882
> Kekulestr. 7, 12489 Berlin, Germany       Fax: +49 (30) 6392 1805
>



More information about the Coin-discuss mailing list