Re: Remote fork() and Parallel programming

Rik van Riel (H.H.vanRiel@phys.uu.nl)
Mon, 15 Jun 1998 12:02:21 +0200 (MET DST)


On Mon, 15 Jun 1998 mshar@vax.ipm.ac.ir wrote:

> The only solution is to refrain from transfering data over a network, but
> some people don't think this is a good idea. Now the main problem is the

On the contrary, this is the best idea there is when you want
to get performance. Just look at the rc5 people (distributed.net),
they manage to scale their project up to 10k nodes with minimal
communication.

Of course, not all projects lend themselves to this efficiency,
but it is a good thing to aim for...

> programming interface: Message passing is no where comparable to DSM in
> ease of use.

MPI can be just as easy as non-coherent DSM. Shared memory is
easy because it's _one_ piece of memory. On the network things
will get duplicated anyway, so things get more and more tricky.
Then MPI will/might be a much better interface.

Rik.
+-------------------------------------------------------------------+
| Linux memory management tour guide. H.H.vanRiel@phys.uu.nl |
| Scouting Vries cubscout leader. http://www.phys.uu.nl/~riel/ |
+-------------------------------------------------------------------+

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu