Newsgroups: comp.parallel.mpi
From: bds@rzg.mpg.de (Bruce Scott TOK)
Subject: Re: blocking nonblocking
Organization: Rechenzentrum der Max-Planck-Gesellschaft in Garching
Date: 11 Jun 1998 16:47:58 +0200
Message-ID: <6loqmuINNabh@s4bds.rzg.mpg.de>

In article <357E1888.41C6@lhm.mw.tu-muenchen.de>,
Martin Schuster  <schuster@lhm.mw.tu-muenchen.de> wrote:
>Hello there,
>
>just recently switched from PVM to MPI getting a little bit confused...
>
>Here's my problem:
>
>For the exchange of boudary conditions performing parallel CFD
>computations, n processes should communicate without data loss
>or gridlocking in the follwing way

 [...]

There's a book called, MPI: the Complete Reference, which shows you all
the different ways to do this.  Which is best depends on your platform;
on the T3E I found SendRecv the fastest.

It is on the big storehouse at

http://www.mcs.anl.gov/mpi/

cu,
Bruce

-- 
"That's the old ecological tale that explains humans' inability to fully
appreciate global warming.  To wit: if you drop a frog in a pan of hot
water, it jumps out.  If you drop it in a pan of cold water, then turn
the heat up slowly, you can roast it to death."  -- Clive Thompson