None of this talk of Beowulf changes two key things as relating to
what I'm interested in, which are that I really do know what I'm
talking about here, having read their academic papers (as well as
Sprite's, as well as GLUnix's, as well as PAPERS') and that what
they're doing doesn't really relate to what I'm doing beyond the fact
that it's parallelized cluster computing.

On Sat, Jul 15, 2000 at 01:24:30AM -0500, Dave Huang wrote:
> I don't know a whole lot about Beowulf clusters, but from what I've
> heard, none of those are the case. Except for possibly #3, depending on
> what you mean by it. According to the Beowulf FAQ at
> <http://www.dnaco.net/~kragen/beowulf-faq.txt>:
> 
>      17. Can I make a Beowulf out of different kinds of machines --
>      single-processor, dual-processor, 200MHz, 400MHz, etc.?
>      [1999-05-13]
> 
>      Sure.  Splitting up your application optimally gets a little harder
>      but it's not infeasible.
> 
> So it sounds like it _can_ be done, but it might be too much work to be
> worth it.

... which is to say that the client programmer better know the set up
of the cluster and start his processes on machines that can handle
them. The onus is on the client, not on the cluster, to distribute
processes intelligently. Obviously, it's not possible to completely
relieve that, but it is possible to make the cluster a little smarter
about it.

(The point is that you can specify a specific node for each parallelized
process you fork off, so if you want something back, you'd better pick
a fast node yourself rather than letting Beowulf pick for you. PVM
behaves in a similar manner.)

       ~ g r @ eclipsed.net