OT, But interesting

[Sean]

On 6 Aug 2000, Ben Pfaff wrote:

> Sean <picasso@madflower.com> writes:
> 
> > Does AMD use IBM's Copper technology? I know Intel licensed the Copper 
> > Technology with the "coppermine" series, 
> 
> I believe that that is incorrect.  I think that Coppermine still
> uses aluminum interconnects.  At least, that's what Tom's
> hardware was claiming back in October:
> 	http://www.tomshardware.com/cpu/99q4/991025/coppermine-02.html
> Maybe it's been done since then?

Tom is cool as hell. 

Maybe coppermine referred to beating the copper process. If Intel is
upping voltages it really doesnt sound like they are using the copper
process. 

> 
> (When do we move to silver interconnects?  Silver is the element
> with the highest conductivity.)

IBM is working on it, but I think there are two drawbacks to it. Silver is
_really_ hard to work with in a pure form, it is very brittle, and has a
very high melting point. Not to mention its rather expensive. 

The copper technology basically wraps, an insulator around the copper so
when its at .18 microns it doesnt short out the chip. Aluminum doesnt need
this because its not as conductive and its a lot less likely to short
circuit the chip. 

I think you will see a move away from silicon first. Something like what
Cray was doing in the 80's, (forgot the name again, it was like adamantium
but thats not even anything real i dont think) But the chips were _very_
brittle and extremely sensitive to temperature changes which isnt good for
PC usage. It was also very expensive. (but probably not relative to the
millions you pay for the cray..)

Actually I know someone at MSU that was trying to synthesize diamonds and
one of the purposes was to replace silicon. Since it is a pure crystal it
can dissapate heat rapidly and evenly and its not conductive. This allows
you to "stack" layers on top of each other. So you could have a small cube
with as many processors as a Cray that could fit in your Palm Pilot. You
of course could use silver with this and even at a .25 micron die it would
probably still be faster and more efficient then aluminum. 

You will see a few other technologies, like light processors, and chemical
processors which might beat it to the shelf though.