GCC 3.0 Release Criteria

[Toon Moene]

Jonathan Thornburg wrote:

> A couple of comments/suggestions:

> >                              LAPACK Fortran 2.0
>                                               ===
> 
> I'm surprised to see LAPACK 2.0 (dated 30 Sept 1994 according to
> the Netlib web site) specified, when LAPACK 3.0update is available
> (dated 31 Oct 1999).  Is there a deep reason for this?

Jonathan, I'm afraid that's my problem.  I indicated to Mark that I
would come up with complete information for testing (and performance
regression testing) for Fortran past weekend.  I knew I couldn't run
LAPACK 3.0 (the June version) on my 64 Mbyte laptop, so I said that and
then tried to download the October version to see if this had the same
problem (the problem is not really in LAPACK - it is due to the fact
that the libf2c runtime allocates some global variables *after* all the
arrays in the main Fortran program have been laid out.  Because some
LAPACK 3.0 test programs use arrays of 10s of millions of DOUBLE COMPLEX
numbers, they cannot be run on anything less than a 0.5 Gbyte
workstation).  I still have to find out.

To cut things short:  I failed to live up to this promise, so Mark just
published what he had ...

> Since LAPACK performance depends heavily on BLAS performance
> (a lower-level subroutine library called by LAPACK), perhaps the
> release criteria should say explicitly what BLAS versions will be used?
> I would suggest either the reference implementation (very vanilla Fortran),
> or (probably more realistic) the Kagstrom, Ling, and Van Loan GEMM-based
> implementation (still portable Fortran, but tuned for computers with
> caches).  Both are available from the same netlib sites that distribute
> LAPACK.

Hmmm, I never tried anything else than the default BLAS - eeks, more
options to test ...

> Right now there's no mention of what LAPACK is and why it was chosen.
> Perhaps a sentence or two should be added, perhaps along the lines of
> 
>    LAPACK is a large subroutine library for numerical linear algebra.
>    It's available in Fortran 77, Fortran 90, and C++ versions.
>    Particularly in its Fortran 77 version, it's representative of
>    much "number-crunching" code, and should particularly stress GCC's
>    loop optimizations.

Thanks for this characterisation.

Hope this explains it a bit (of course, lack of spare time is an excuse
unworthy of a physicist - we could always create a wormhole ...)

-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://xena.eas.asu.edu/~andy (under construction)