[ExI] gimps

[Eugen Leitl]

On Fri, Jun 22, 2012 at 07:59:15AM -0700, spike wrote:

> While we go on about technological change slowing, note that my own memories
> are still fresh of the time when a typical computer could do only a fiftieth
> of what the typical computer does today.  The computers blew our minds back
> then.  Huh.  You young people don't know how good life is today.

The rate of growth has been slowing down for a while, though. You're lucky
in that yours is a specific numerical application that happens to
profit unusually good from ALU speed.

The clocks effectively stopped doubling in 2003, and it's been not so 
good in general http://www.gotw.ca/publications/concurrency-ddj.htm

In order to continue scaling as long as Moore's going to hold (not
for very much longer, the 2d litho limit is probably around 1-3 nm)
we need different architectures, and people who can deal with asynchronous
shared-nothing distributed (and even nondeterministic) applications.

We'll probably run out of Moore well before the paradigm switches from
multithreaded to the asynchronous nondeterminism.

Sorry to be the perpetual party-pooper. On the positive side, we'll 
need to have architectural needs and 3d volume integration now must come by hook or
by crook.