[ExI] Moore's Law

[Eugen Leitl]

On Wed, Mar 20, 2013 at 10:33:07AM -0700, Adrian Tymes wrote:
> On Wed, Mar 20, 2013 at 10:09 AM, Eugen Leitl <eugen at leitl.org> wrote:
> > There is no fundamental progress in software. The progress in hardware has
> > recently been limited, especially since Moore has ended.
> 
> The data I've seen suggests that Moore's Law continues through the

http://www.semiwiki.com/forum/content/1388-scariest-graph-i-ve-seen-recently.html

Published on 07-01-2012 02:00 PM

"This shows the cost for a given piece of functionality (namely a million gates) in the current process generation and looking out to 20nm and 14nm. It is flat (actually perhaps getting worse). This might not matter too much for Intel's server business since those have such high margins that they can probably live with a price that doesn't come down as much as it has done historically. And they can make real money by putting more and more onto a chip. But it is terrible for businesses like mobile computing that don't live on the bleeding edge of the maximum number of transistors on a chip. If you are not filling up your 28nm die and a 20nm die costs just the same (and is much harder to design) why bother? Just design a bigger 28nm die (there may be some power savings but even that is dubious since leakage is typically an increasing challenge).

If this graph remains the case, then Moore's Law carries on in the technical sense that you can put twice as many transistors on your chip if you can think of something clever to do with them and can find a way to keep enough of them powered on. But it means there is no longer an economic driver to move to a new process unless you have run out of space on the old one.
"

> present day - at least for actually released hardware.  There is no

Yeah, I like to call things before most people notice.

> end of reasons to believe that the next hardware won't quite keep up
> the improvement, that we have reached fundamental physical limits

Not fundamental physical limits. Economic limits, since Moore is
about *affordable* transistors. 

> or the like - and yet, the hardware keeps improving, maintaining its

It does not seem you've been looking at architecture trends even while
Moore was on-track. 

> net overall increase pace by other means.  Often this is done by
> incorporating design tricks that are not technically impacted by said

My point is that the architecture changes had no imagination. It didn't
have to, while the amount of widgets quadrupled. Now, there's an 
incentive to rearrange the real estate in novel ways.

> physical limits.  For instance, even if we were at the limit for the
> minimum number of atoms one could use for memory, there is now
> research in encoding bits into individual electrons, allowing for

We're not talking research, which is 20-30 years away from
production. We're talking about production.

> multiple bits per atom...and there may be ways to do multiple bits
> per electron.

Nope. You need about nm^3 for an addressable bit.
 
> http://en.wikipedia.org/wiki/Moore%27s_law has data that shows it
> continued at least through 2011.  So if Moore's Law has ended, one

We've got March of 2013. 2011 is ancient history.

> would need to show that with data from 2011-2013,..and as late

You'll see the saturation setting in over the next 2-5 years.
It will be completely obvious, in the hindsight.

> 1990s data from that graph shows, the long-term progress continues
> even if there are slowdowns for a few years.
> 
> So - I'd say the data rather firmly shows that Moore's Law has not
> ended.

I say rather firmly that you haven't been been looking at the data.