Subject: Re: what to do with an idle VAX...
To: John <john@sixgirls.org>
From: Brian Chase <bdc@world.std.com>
List: port-vax
Date: 06/18/2001 23:41:26
On Mon, 18 Jun 2001, John wrote:
> > > Too bad we don't have a distributed.net client - I've been running dnet on
> > > all of my m68k NetBSD systems, and there's a separate grouping for m68k
> > > NetBSD (and 90% of all of the work is me!):
> > > http://stats.distributed.net/ogr-25/platformlist.php3?view=toc
>
> > I would rather have a VAX/VMS client for SETI@Home. The mission is more
> > appealing to me than that of distributed.net.
>
> They don't have clients for ANY alternative OS or CPU. They're also quite
> vocal about making people understand that it's not worth their time
> because alternative CPUs and OSs don't do enough work...
>
> At least OGR has some good mathematical applications. That's why I run it
> on all of my machines.
The distributed.net clients and the SETI@Home clients for the VAX (under
VMS and NetBSD) have been investigated in the past. The main reasons that
they aren't being supported are...
[1] The VAX doesn't support IEEE floating point. This throws a wrench
into the SETI@Home efforts.
[2] The turn-around on work units would be so slow on the VAX, that units
wouldn't be completed in the allowed time window.
It's true that they could expand the time window for unit completion, but
given the slow speed of most VAXen, the distributed.net speed freaks just
aren't interested. I believe the IEEE issues with SETI@Home are valid.
We could write an IEEE compatible library, but performance would be slow.
It would be nice to gather up a list of NetBSD/vax friendly distributed
computing efforts. Given that everyone else is so busy with SETI@Home and
distributed.net stuff, the VAX community might be able to help some of the
more neglected projects. :-)
The http://www.mersenne.org/ site has a number of projects listed. Some
of them would perhaps want our cycles.
-brian.