Breaking News
News & Analysis

Samsung Shows 14nm Chip

10/2/2014 07:40 AM EDT
7 comments
NO RATINGS
1 saves
More Related Links
View Comments: Newest First | Oldest First | Threaded View
goafrit
User Rank
Author
Re: 0.7V potential
goafrit   10/4/2014 1:39:40 PM
NO RATINGS
>> There has been number if studies that seem to say that the energy consumption is the lowest around 0.4V...but in practice as you say variability is the key

That is the key point. Once you pass the Vth, you have a chip firing. If you stay there, you may lose performance. The benchmark on operating voltage must include  performance and variability to make any sense.

goafrit
User Rank
Author
Re: 0.7V potential
goafrit   10/4/2014 1:37:55 PM
NO RATINGS
>> Someone at the booth today told me the process has potential to drive chips at voltages as low as 0.7V.

That is a wake-up call for Intel that no one has this monopoly of new ideas i this industry. Intel has been putting money in this area. It is good to see Samsung coming big on board.

Gondalf
User Rank
Author
Re: Who will win?
Gondalf   10/3/2014 4:04:32 PM
NO RATINGS
I think you'll find only closed mouths around Rick :).

I am a little bored about the lack of informations about yields because this was Intel main issue in last quarters. Obviously this thing is even more interesting to know for Foundries, thier products rarely have an high ASP.

About your "who will win"; i think this is not so important in present foundry landscape. TSMC has the volume, Samsung and GloFo nope.

At the end of the day TSMC care nothing to be the "first", it is the only foundry around the world capable to ship the real volume right now.

The real questions:

1) the customers are convinced of the performance advantage of 16nmFF over vanilla 20nm?? The claimed +11% in clock speed of a A57 and a tiny (my estimation) 10/15% advantage in power consumption in complex compute parts at the same clock speed, are enough to justify a new mask set and a new circuitry layout??

2) the cost of single transistor is down for real?

IMO 20nm eat a lot of 14/16nm future. 

krisi
User Rank
Author
Re: 0.7V potential
krisi   10/2/2014 7:57:40 PM
NO RATINGS
There has been number if studies that seem to say that the energy consumption is the lowest around 0.4V...but in practice as you say variability is the key

AKH0
User Rank
Author
Re: 0.7V potential
AKH0   10/2/2014 5:11:24 PM
NO RATINGS
You can operate any CMOS technology at 0.7V or even smaller. Performance of course drops so does power. The real barrier to low-Vdd operation is device variability. The question is if FinFET is the answer to variability.

rick merritt
User Rank
Author
Who will win?
rick merritt   10/2/2014 2:24:54 PM
NO RATINGS
I'd love to hear informed opinion of who is ahead and why in the 14/16 FinFET race.

rick merritt
User Rank
Author
0.7V potential
rick merritt   10/2/2014 2:23:52 PM
NO RATINGS
Someone at the booth today told me the process has potential to drive chips at voltages as low as 0.7V.

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed