Re: MPI

[Alex Verstak]

I cannot give you an example, but I can explain why they wrote
it in the standard this way.  MPI/C is a low-level API for a
low-level language, so this requirement is not unreasonable.
A few parallel systems at the time passed start-up information
as arguments, so the MPI folks decided to make life easier for
those systems.  Standards are not meant to create good designs;
they merely codify the existing practice.  (As a side note, MPI
is one of the cleanest standards in parallel computing.)

I already explained how MPICH uses this feature.  I don't know
about other MPI implementations, but experience suggests that
any provision of a standard, no matter how weird, will be used
by some implementation just because it can be.

=alex


On Fri, 2 Feb 2001, John W. Eaton wrote:

> On  2-Feb-2001, Alex Verstak <address@hidden> wrote:
> 
> | A quote from the MPI Standard:
> | `An MPI implementation is free to require that the arguments in the C
> | binding must be the arguments to main.'
> 
> This makes no sense to me.  Can you (or anyone) give me an example of
> a case where it would matter?
> 
> jwe
> 



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------