[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: MPI
From: |
Alex Verstak |
Subject: |
Re: MPI |
Date: |
Sun, 11 Feb 2001 05:49:19 -0500 (EST) |
On Sat, 3 Feb 2001, John W. Eaton wrote:
> I took the time to look up the MPI standard, and here is what I found at
> http://www-unix.mcs.anl.gov/mpi/mpi-standard/mpi-report-2.0/node43.htm#Node43:
<snip argc=NULL, argv=NULL comment>
> So I don't think it matters whether you have the exact argv/argc that
> are passed to main.
MPI 2.0 is not widely supported yet. The latest and the greatest
MPICH 1.2.1 crashes on MPI_Init(NULL, NULL).
> (BTW, I think there is a slight error in the
> standard, since an argc == NULL doesn't make sense.)
It's a pointer to argc.
> However, at http://www-unix.mcs.anl.gov/mpi/www/www3/MPI_Init.html I
> see that
>
> The MPI standard does not say what a program can do before an MPI_INIT
> or after an MPI_FINALIZE. In the MPICH implementation, you should do
> as little as possible. In particular, avoid anything that changes the
> external state of the program, such as opening files, reading standard
> input or writing to standard output.
>
> But I suspect that this still doesn't mean that Octave must call
> MPI_Init before doing anything, or that it must be done in main(), or
> that it can't go in a .oct file, though you may need to call MPI_Init
> before doing any other operations in your program.
For that matter, MPI doesn't say that file i/o is available any
time, not just before MPI_Init. Chances are, Octave won't run on
machines where it matters anyway. ;-)
Bottom line, I got the .oct file to work with MPICH on a
Linux/Myrinet cluster and with LAM on a cluster of workstations.
No changes to Octave necessary. I'll test it on a few other
configurations before releasing the source (may take a few weeks).
=alex
-------------------------------------------------------------
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
-------------------------------------------------------------
- MPI, Alex Verstak, 2001/02/01
- MPI, John W. Eaton, 2001/02/01
- Re: MPI, Alex Verstak, 2001/02/02
- Re: MPI, John W. Eaton, 2001/02/02
- Re: MPI, Alex Verstak, 2001/02/02
- Re: MPI, John W. Eaton, 2001/02/02
- Re: MPI, Alex Verstak, 2001/02/02
- Re: MPI, John W. Eaton, 2001/02/03
- Re: MPI, Alex Verstak, 2001/02/03
- Re: MPI, John W. Eaton, 2001/02/03
- Re: MPI,
Alex Verstak <=
Re: MPI, Buenner, Martin, 2001/02/02
Re: MPI, Andy Jacobson, 2001/02/02