Re: alternate simplified proposal #1 (fwd)

Richard Frost (frost@sdsc.edu)
Mon, 10 Jun 1996 12:23:00 -0700 (PDT)

In a few discussion groups on Friday, it was proposed that the
"info" structure being proposed in mpi-dynamic become an explicit
MPI opaque object (MPI_INFO).

If this occurs, then a general MPI mechanism for constructing and
manipulating {key, value} pairs could be used to specify info (e.g.,
"hints") about the execution environment.

I suppose this would go in the miscellany chapter. Any volunteers
to write it?

- Richard

---------- Forwarded message ----------
Date: Mon, 10 Jun 1996 09:45:36 -0700 (PDT)
From: Parkson Wong <parkson@nas.nasa.gov>
To: Yoonho Park <yoonho@sina.hpc.uh.edu>
Cc: mpi-io@mcs.anl.gov
Subject: Re: alternate simplified proposal #1

>
> > ** In contrast to the existing proposal (10.7), implementation
> > directives may NOT be given in the filename, but rather
> > are specified in the "info" argument.
>
> Richard, can you provide some examples here?
>

Info (the new word for hints) is just a hint. We can't expect all
implementation to do something with it, so moving the directives from
the filename to info has quite different semantics.

In the original spec, if you open piofs:/foo, and that system does not
support piofs, an error will be returned. (It may not have specified
that way, but it is the intent of the spec.) If this is in the info, it
is just ignored.

-- parkson

--
Parkson Wong			Address: Numerical Aerodynamic Simulation
MRJ, Inc.				 NASA Ames Research Center M/S 258-6
Supercomputer Applications Segment	 Moffett Field, CA  94035-1000
e-mail: parkson@nas.nasa.gov	Phone: (415)604-3988	Fax: (415)966-8669