New Energy Efficent Micro-Chip

[Nflight]

http://www.cellular-news.com/story/29048.php?source=rss

Looks like everything is really getting smaller all the time, you read about a development in one part of the country and within days someone from somewhere else makes it even smaller still. And we haven't even moved out of the country, where the world is downsizing and removing barriers all the time. We all live in an amazing society!

[AMDave]

The numbers in that journalist's article don't quite add up. 10 times more efficient would be a 0.1 volt draw down from 1.0 volt. 90% reduction leaving a 10% draw. 10% going into 100% 10 times.

The researchers are promoting a draw of 0.3. By my reckoning that is a 70% reduction in quiet-mode power consumption
30% draw goes into 100% 3.3r times. So we have something that is actually just over 3 times more efficient.

None the less, the subject is a substantial improvement and is highly remarkable.

Hopefully the Proof of Concept path is not to expensive for manufacturing designers to take up. Getting a laptop or phone to last 3 times longer in quiet-mode would reduce the number of recharges and extend the life-cycle of the battery as well. That takes a load off the battery manufacturers and reduces the horrible wastes that those processes produce.

Given that they have also had to implement changes to make the silicon paths more reliable, it is likely they may have found yet another aspect that will extend to push Moore's law further. We already have silicon-polymer based dies so the potential enhancements could get into the process sooner than the new non-silicon proof-of-concepts go to manufacturing. My guess is the could fill in the gap nicely. Silicon shares, anybody?

Back-slaps and handshakes all round for the MIT Prof's if it all comes to fruition.

[AMDave]

dig dig...
The claim of 10 times was made by David Chandler (MIT) here in the original
http://web.mit.edu/newsoffice/2008/techtalk52-15.pdf
That seems to be the article that has been syndicated all over the news-wire with that claim.

I'd prefer to read the facts in the paper by the participants...

the department announcement was made here
http://www.eecs.mit.edu/cgi-bin/announcements.cgi?page=2008/data/35.dat
and it also states up to 10 times, so they are definitely referring to some greater improvements that they likely shared with those at the IEEE conference.

The paper was to be delivered on Feb-5 here: http://128.100.10.145/isscc/
I can't find the paper in the IEEE archive yet or here: http://mtlweb.mit.edu/~anantha/publications.html (http://mtlweb.mit.edu/%7Eanantha/publications.html)

Sounds very promising indeed, though.