Path: utzoo!attcan!uunet!lll-winken!brutus.cs.uiuc.edu!zaphod.mps.ohio-state.edu!think!yale!cs.yale.edu!zenith-steven
From: zenith-steven@cs.yale.edu (Steven Ericsson Zenith)
Newsgroups: comp.sys.transputer
Subject: Re: Re 64K Node Machine
Message-ID: <19312@cs.yale.edu>
Date: 16 Mar 90 15:31:26 GMT
References: <1828.9003152120@prg.oxford.ac.uk>
Sender: news@cs.yale.edu
Reply-To: zenith-steven@cs.yale.edu (Steven Ericsson Zenith)
Organization: Yale University Computer Science Dept., New Haven, CT  06520-2158
Lines: 51

In article <1828.9003152120@prg.oxford.ac.uk>,
HALLAM@physics.oxford.ac.uk ("Phillip M. Hallam-Baker") writes:
> 
> At Southampton there is a big(ish) array of 1260 T212's without external
> RAM (called `Deep Thought' as each board has 42 nodes...). There was talk
> of upgrading the RAM to 64K per node at one time maybe someone garbbled the
> message somwhere? - either from that machine or a similar T2 engine.

AN interesting machine, born of a fortuitous error. 
Someone bonded a large number of devices (either 90 
or 180 deg rotation) incorrectly in their packages. 
Guy Harriman had special boards made to accomadate 
the error.

> Quite
> what can be done with a 64K NODE machine I don't know - surely the link
> speed would start to be a problem? If not that how about the physical cooling
> /mounting engineering type problems? Where to get the 256 Gigabytes of RAM to
> make it worth while? Sounds like a fun project! 

Now here's an interesting question. What would the 
characteristics of such a machine be? Remember, there
is only 2kbytes per node, fixed configuration. 
I figure the only useful way to program such a machine
would be of the "load problem, compute, unload solution"
variety, perhaps with an exchange with nearest nieghbour
in there somewhere. But then I'm sure for most problems 
2K is just not going to be enough. Anyone know better?

>- Anybody out there want to
> write me a blank cheque to build one?

Make that cheque out to the two of us!

> 
> 			Phillip Hallam-Baker
> 
> 
> 				Oxford University Nuclear Physics
> 				ZEUS group 
> 
> "You don't have to write in FORTRAN to work here - but it helps."

"You don't have to write Occam to work here - and they'd rather you
didn't." sez.
--
.                                                                        .
Steven Ericsson Zenith              *            email: zenith@cs.yale.edu
Department of Computer Science      |            voice: (203) 432 1278
Yale University 51 Prospect Street New Haven CT 06520 USA.
     "All can know beauty as beauty only because there is ugliness"