[rescue] Nano-tech storage (was: Tape drive woes)
Francisco Javier Mesa-Martinez
lefa at ucsc.edu
Wed Sep 15 00:05:23 CDT 2004
On Tue, 14 Sep 2004, Dan Duncan wrote:
> On Tue, 14 Sep 2004, Francisco Javier Mesa-Martinez wrote:
> > If it is stuff that you can mix with the in-die logic then it would be
> > interesting. If you have to access off chip, it doesn't matter how fast
> > you can access stuff, since you are still bitten by the wire delays and
> > pin limitations :(.
> Yeah, you're speed of light limited to what, 0.3mm per picosecond?
Speed of light is not your main limitation, electromagnetism is. Frequency
is the main limiter when it comes to the actual length of your features,
you could do fancy differential pipelined signaling (which is all the rage
right now, with stuff like openTransport) but you still have reflection,
inductance and most important crosstalk. You also have the bandwidth
constrains for each pad/pin/contact technology. At least with current
board level state of the art. And since dollars dictate technology, it
seems that we'll be stuck with CMOS and traces for a long time :(.
More information about the rescue