|
|
Glas :Re: [glas] summary of fonctionnalities |
From: Karl Meerbergen (Karl.Meerbergen_at_[hidden])
Date: 2005-02-12 12:23:23
On Wednesday 09 February 2005 23:41, Simon Perreault wrote:
> On February 9, 2005 13:59, C J Kenneth Tan -- OptimaNumerics wrote:
> > It would make sense to have code that just translates to BLAS calls,
> > and use BLAS code from an implementation of BLAS. While it is easy to
> > get a working implementation of BLAS code, it is actually not always
> > easy to get good BLAS performance. So it would make sense to have
> > GLAS code just calls BLAS code, to take advantage of tuned BLAS
> > implementations.
>
> Sorry to interrupt here, but I don't think GLAS is supposed to be a wrapper
> for BLAS or LAPACK, is it?
Of course not. Neither is it a wrapper for any other software. But it is a
requirement, which makes sense: rewriting LAPACK cannot be the goal for GLAS
(at least not for me that is) and if the performance of BLAS is better than
anything else, it would be stupid not to use it.
Clearly, the development of efficient C++ software for matrix - matrix product
for example, is important and therefore it is another requirement of GLAS.
Also note that the BLAS do not provide (nice) software for the sparse case
(unless the performance has improved). Can this be confirmed by someone? Also
structured matrices are not covered by BLAS. Nor is mixed value_type
computations. In other words, GLAS will be more complete than BLAS.
Karl
-- ============================================== Look at our unique training program and Register on-line at http://www.fft.be/?id=35 ---------------------------------------------- Karl Meerbergen Free Field Technologies 16 place de l'Université B-1348 Louvain-la-Neuve - BELGIUM Company Phone: +32 10 45 12 26 Company Fax: +32 10 45 46 26 Mobile Phone: +32 474 26 66 59 Home Phone: +32 2 306 38 10 mailto:Karl.Meerbergen_at_[hidden] http://www.fft.be ==============================================