Beowulf in a Box (fwd)

[Miguel Barreiro Paz]

> > 	It's neat, but with just 32 bits for addressing we're getting a
> > bit tight of space on x86.
> 
> 32 CPUs each with 64MB of memory is 2G of memory - plenty of space
> left over ;)

	There are boards with dual PCI buses out there, and (clone) x86
systems with 1GB of RAM. Aren't we pushing the limits a bit too much? :)

> Consider a slight alteration to what I proposed though: say each
> processor has 1G of memory associated with it, but only map 16MB or 32MB 
> of that processor's memory into the address space of the host.  This
> would be the shared memory for each processor - the rest would be
> for local computation only.

	Yes, that's not so elegant but would work, I suppose. Or you could
simply access the host RAM for many algorithms. I'm afraid we are talking
about vapor software on (currently) vapor hardware, let's get some
hardware done and then work on it.

Regards,

Miguel