[ExI] The Jaguar Supercomputer

Eugen Leitl eugen at leitl.org
Tue Nov 17 10:59:56 UTC 2009


On Tue, Nov 17, 2009 at 11:43:38AM +0100, Stefano Vaj wrote:

> One wonders what the closest "wall", making a breakthrough necessary to go
> on, might lie...

That is easy, smooth takeover into 3d integration when 2d lithography
will run out of steam (when is anyone's guess, though
http://en.wikipedia.org/wiki/11_nanometer appears as good a
candidate as any).

Of course you can still make architectural advances, go stacked
silicon with TSV http://en.wikipedia.org/wiki/Through-silicon_via
go WSI http://en.wikipedia.org/wiki/Wafer-scale_integration
but your only next chance is to abandon flatland, go 3d
(whether by iterative layer deposition, or self-assembly of
bulk), go spintronics, reversible computer and similiar
technologies which are breakthrough-driven, and hence cannot
be extrapolated as a smooth semilog plot -- unless you're
very, very lucky, and one technology so happens to be ready
to smoothly pass over the torch.
 
> In the meantime, let us enjoy the steady performance increases... :-)

Absolutely. It's not going fast enough by far; we need a 50 kUSD
supercomputer under everybody's desks yesterday.

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE



More information about the extropy-chat mailing list