main

Intel Says 45nm Process On Track

stncttr908   on 25 January 2006 - 19:57 · 30 comments & 8211 views

Advertisement (Why?)
Intel, the world's largest microprocessor manufacturer, has created test chips made using the 45nm fabrication process and expects to see processors, flash memory, and other products based on the technology available in the second half of 2007, according to director of process architecture and integration Mike Bohr.

The test chips are static SRAM modules containing 153 megabits of memory. By comparison, a chip similar in size made with a 130nm process in 2000 contained only 18 megabits. The nanometer measurement is a billionth of a meter, and refers to the size of features on the processor, such as transistors. By comparison, a human hair is between 60 and 90 microns, or millionths of a meter, in diameter.

By shrinking the transistors, energy consumption and production costs are reduced while higher performance yields are obtainable.  While these are only test chips, it is an important indication that 45nm tech is still on track as forecasted.

News source: CNET News.com




Post a comment · Send to friend Comments · There are 30 additional comments
#1 ShadowPHP on 25 Jan 2006 - 19:59
YaY! Go NANOMETERZ!
(3 replies) #2 BanneD on 25 Jan 2006 - 20:21
that's pretty impressive, wondering when will AMD announce something like that
#2.1 phun8890 on 25 Jan 2006 - 21:07
I think AMD is buying 45nm from IBM and will in the process of developing in 2008. Not sure...
#2.2 VikingStorm on 26 Jan 2006 - 00:37
I think AMD and IBM are working together for that (not buying from each other).
#2.3 tiagosilva29 on 26 Jan 2006 - 17:05
Either way, it's going to kick ass.
(2 replies) #3 Lasker on 25 Jan 2006 - 20:23
45nm is a complicate process, is more small than a human hair
#3.1 MNS on 25 Jan 2006 - 22:57
transistors have been smaller than a human hair for decades now.
#3.2 vacs on 26 Jan 2006 - 18:46
lol Lasker, good one

Do you know why it is called nanotechnology? hehe
(4 replies) #4 Lee McDermott on 25 Jan 2006 - 20:27
Are there any performance gains to be had by reducing the size? What are the benefits?
#4.1 Jason on 25 Jan 2006 - 20:33
Less heat and less power needed and more chips per wafer thus reducing manufacturing costs.
#4.2 stncttr908 on 25 Jan 2006 - 20:42
Probably best if that were including in the article, eh?
#4.3 sphbecker on 26 Jan 2006 - 16:56
I thought it was common knowledge that reducing the size of a chip reduces power and heat, and increases performance.

I think that is because the smaller chips require electrical signals to travel shorter distances, which means you can use lower the amperage, which in turn means smaller transistors and wires. Less heat because you are using less power and all power is converted to heat at some point. The shorter distances also allow singles to reach their destination more quickly; the speed of light is a big factor in someone as performance critical as a processor.

I'm sure its much more complex then that, but that is probably ruffly why smaller is better.
#4.4 Blackice on 26 Jan 2006 - 19:29
sph: you're right, but electrons don't travel at the speed of light (they can come quite close, but then you get into relativistic mechanics, where nothing can travel at the speed of light).

Another good thing aobut going to 45nm chips is that more chips can be made per silicon wafer, meaning less expensive chips (theoretically).
(5 replies) #5 Intelman on 25 Jan 2006 - 20:55
How small can they go :p Wouldn't there be a limit?
#5.1 Jack31081 on 25 Jan 2006 - 21:09
well, zero is certainly a limit
#5.2 MNS on 25 Jan 2006 - 23:05
indeed.

what happens after they pack several million transistors into a single nanometer? which, as i see it, is bound to happen in the next 10 years..

talk about a complete redesign of the computing world, which by that time we will be even more immersed into, making it that much harder to transition..

they will need to think of something in a more abstract way, since we'll already hit the limit in the physical realm.. either that, or invent atom-sized robots that perform the building tasks of these sub-atomic transistors.. even then, you'll still have the atoms as a limit.. we need to come up with something entirely different.. i can't wait!
#5.3 Croquant on 25 Jan 2006 - 23:09
The limit isn't zero, but it might as well be. We're getting to the point where soon the only way to achive greater computational speeds will be to either go to superconducting materials (most likely supercooled ceramics could be used), or abandon electrons as our information cariers altogether and switch over to photonic (light-based) processors. Of the two, the photonic-based computer seems to hold the most practical worth, as some successfull work has already been done in this regard (although it's all a long way from being out of the experimental stage).
#5.4 tiwaris on 26 Jan 2006 - 01:59
I guess, nano-meter is already the limit not because the tech cannot advance, but they are already working at the order of molecule-size (10 nm). To go beyound the nm scale, a totally new concept will be required which is not hindered by size of molecules and atoms being used.
#5.5 Blackice on 26 Jan 2006 - 19:45
using light instead of electrons seems bad to me. Ok, it moves faster (ok then, a lot faster), but I feel all the light emitter/detectors on a chip would lead to bigger computers... I'm guessing.

But light certainly does seem to be what's going to happen (they'll figure a way to shrink it all). That means they'll put it everywhere - in all the cables and that. Actually, I'm surprised we don't already use optical IDE cables. It's not like much would need to change (unless I'm ignorant).

But that's all going to take a long time. Superconductivity is probably what we're going to see first... and damn that will be cool!
#6 Pete Zaria on 25 Jan 2006 - 21:27
It just never ceases to amaze me.... 50 years ago, our best supercomputers were using vacuum tubes as transistors, each one was 3 feet tall. Now, we're packing what, 45 billion transistors into a 1-meter area? My laptop has more technology in it than the space shuttle had on it 10 years ago. Where will we be in 10 more years?

As to what good does 45nm do...
A) smaller components mean packing more "stuff" into a smaller space
B) probably lower manufacturing costs, once R&D is complete
C) potentially lower heat production and power consumption, which is always a plus
D) basically, smaller, faster, cooler chips that run on less juice than their predecessors.

Very cool stuff. Here's hoping AMD is right behind them.

Peace,
Pete Zaria.
(1 reply) #7 ECEGatorTuro on 25 Jan 2006 - 23:04
This is the process that the new Intel fab in Arizona will be running. It is the largest contruction site in Arizona right now and it's an amazing site. Can't wait to start working in the new Fab!
#7.1 Steven on 27 Jan 2006 - 09:07
I'm sure they'll have no problem finish it as the skies refuse the cry for us. 101 days without rain.. :bangs head:

YES I KNOW I LIVE IN A DESERT BUT THIS IS REEEEEEEEEEEEDICULOUS!
(3 replies) #8 OfF3nSiV3 on 26 Jan 2006 - 01:30
what speeds will we get with 45nm?
speeds have been stalled for a long time now..
#8.1 Andre on 26 Jan 2006 - 02:20
It's not about speed anymore.
#8.2 OfF3nSiV3 on 26 Jan 2006 - 02:26
hmm it's about what then?
#8.3 ECEGatorTuro on 26 Jan 2006 - 03:04
It's about redesigning the architectures to do more per clock cycle and to do it more efficiently and with less heat dissapation.
#9 LTD on 26 Jan 2006 - 06:51
I'm still pretty happy with my Barton-core Athlon 3000+

LOL. ;-)
#10 Suren on 26 Jan 2006 - 13:33
Thanks for your help!

Commenting has either been disabled on this article or you are not logged in. Click here to login or register, its free!

Note: Anonymous commenting is disabled in order to keep the quality of responses to a high standard.

Advertisement (Why?)