Re: Apple II suicide--a sermon.(parallel II's?)

[jd-wong@uxa.cso.uiuc.edu (Jim Wong)]

ronlouie@u.washington.edu (Ron Louie) writes:

>Naive question from a non-computist:  Could one create some kind of 
>parallel "supercomputer" using ole' Apple II's?  I read (in the New 
>Yorker a few years ago) about enterprising mathematicians wiring up some 
>pre-486 PC's in their NYC apartment to compute Pi to record levels.  
>Their Rube Goldberg computer was supposed to rival Crays and Japanese 
>supercomputers.  So.....if we wired up all the elementary school II's out 
>there, can we use them to do the latest Top Quark and Fermat's theorm 
>calculations?

The II has several limitations that would make building a high performance,
distributed memory parallel machine out of them difficult.

First, it has fairly limited memory; it's hard to work on large problems
when you've only got 64K - OS overhead on each processor.

Second, the II has very limited I/O bandwidth.  This weakness is
exacerbated by the memory limitations I mentioned above: with only
a small amount of memory, there's going to be a lot of communication
going on, both between processors and memory and processors and
other processors.

Third, the 6502 isn't a very good chip for doing the kind of floating
point intensive computations one usually performs on a supercomputer.

All that aside, I'm sure you could do some interesting things with
an array of II's -- it's just a question of whether it's worth the
effort to put together an interconnection network and write a
Fortran or C compiler to support such a beast.
-- 
Jim Wong (jd-wong@uiuc.edu)