Re: [MLUG - DISCUSSION] I.B.M. Advance Connects Layers of Tiny Wafers

[Jonathan King <EMAIL:PROTECTED>]

On Mon, 11 Nov 2002, Mike Miller wrote:

> FYI...
> 
> I.B.M. Advance Connects Layers of Tiny Wafers
> from The New York Times

[snip]
 
> One advantage of the technique would be to interconnect separate
> layers directly at thousands or even hundreds of thousands of
> points. Computer speeds are currently limited by the bottleneck
> of getting data on and off individual chips. Increasing the
> number of connections may sharply increase communication speeds.

Wow.  This is going to be a pretty major advance.  For a long time,
my big concern was how I/O-starved CPUs were getting.  Part of the
answer there would be tremendously large caches.. . but then there 
was the issue of bandwidth between cache and processor.  This should 
help that quite nicely, while decreasing the space needed for lots 
of other support chips that can now be integrated "vertically" as it 
were.  This is definitely major-league cool.

jking

--
To unsubscribe, go to http://mlug.missouri.edu/members/edit.php

Archives are available at http://mlug.missouri.edu/list-archives/