backtop


Print 59 comment(s) - last by Justin Time.. on Mar 31 at 3:34 AM

MIT researchers have devised a graphene signal multiplier that produces a clean signal with low power

The computer industry is involved in a never-ending quest for the highest possible performance from computer components and other electrical devices. Over the years, there have been many advances that have boosted the computational power of computer systems from increasing the number of transistors to adding multiple processing cores.

Researchers across the world are hard at work on building microprocessors and other electrical components using a material discovered in 2004 called graphene. A group of researchers at MIT announced on March 19 that new findings made could lead to much faster microprocessors in the future. The new findings could lead to cell phones and other communications equipment that can transmit data much faster than devices available today.

The researchers at MIT have built an experimental frequency multiplier made from graphene. The multiplier is capable of taking an incoming electronic signal at a specific frequency and producing an output signal that it a multiple of the original frequency. One example of a use for the graphene frequency multiplier is inside a microprocessor to determine the clock speed of a CPU.

Frequency multipliers are used widely today according to MIT, but the difference between the multipliers we have today and the graphene multiplier is one of signal noise. Today's multipliers produce noisy signals that require filtration and consume lots of power. The new graphene frequency multiplier consists of a single transistor and produces a clean output signal in a highly efficient manner.

MIT assistant professor in the Department of Electrical Engineering and Computer Science Tomas Palacios said in a statement, "In electronics, we're always trying to increase the frequency. It's very difficult to generate high frequencies above 4 or 5 gigahertz."

Palacios says that the graphene multiplier could lead to systems capable of operating in the 500 to 1,000 GHz range.

Palacios continued saying, "Researchers have been trying to find uses for this material since its discovery in 2004. I believe this application will have tremendous implications in high-frequency communications and electronics."

The graphene multiplier is still at the laboratory stage but the researchers believe that practical working systems using the graphene technology could be available in one to two years. The findings are being reported in a paper to be published in the May issue of Electron Device Letters.

The MIT researcher's multiplier is far from the only graphene device being tested in labs around the world. A group of researchers from Rice University created a new type of memory from graphene in December of 2008. The discovery could lead to graphene storage arrays in the future.

One of the key findings that may help push research into graphene based electronic devices was a discovery made in January 2009 by a group of researchers from the Rensselaer Polytechnic Institute. The researchers discovered an easy way to control whether graphene is a semiconductor or a metal. Graphene in pure semiconductor form is required for use in computer processors and other components. The researchers explain that even a tiny section of graphene in metallic state along with mostly semiconducting graphene will destroy the semi-conductive properties of the graphene sheet.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Odd
By Oregonian2 on 3/24/2009 12:33:13 PM , Rating: 4
I think the basis of the article is funny. I don't think the fundamental limiting problem in CPU design is the ability to make a PLL with a suitably high output frequency. Don't think that's anywhere near the biggest problem.

If they were able to make a baseball bat that wouldn't break even if it were used to hit a baseball 100 Miles, would major league players start hitting them that far?




RE: Odd
By Alphafox78 on 3/24/2009 12:36:17 PM , Rating: 3
yes.


RE: Odd
By JasonMick (blog) on 3/24/2009 1:35:25 PM , Rating: 2
But only if they took received their pre-workout "flaxseed cream" rubdown.


RE: Odd
By paydirt on 3/26/2009 10:48:06 AM , Rating: 2
:) most computers still have much more processor power than most folks need.

It would be cool to get more power for scientific purposes such as computational biology.


RE: Odd
By hlper on 3/24/2009 1:18:55 PM , Rating: 5
quote:
If they were able to make a baseball bat that wouldn't break even if it were used to hit a baseball 100 Miles, would major league players start hitting them that far?


If we start making next generation steroids from graphene, then yes ;-)


RE: Odd
By bighairycamel on 3/24/2009 3:46:51 PM , Rating: 2
Yah I agree.

One major underlying problem with a digital signal is skew rate. A digital signal doesn't instantaneously switch from low to high or vice versa, even in a single transistor. It has a finite time it takes to "ramp" up to the necessary voltage. I think now a CPU reads 1.2V or higher as a binary 1 (correct me if I'm wrong) and 285mV or lower as a 0 (again, may not be up to date). At 1000GHz a single bit probably wouldn't reach 1.2V before the next one has been received.


RE: Odd
By wifiwolf on 3/24/2009 10:12:10 PM , Rating: 4
I agree with you 2 but you're not seeing all the variables.
Some time ago we used 0v to 5v to signal '0' and '1'.
Now it's narrowed.
I'm sure we can narrow it much more when we get to need it.
There is a Vdd and vcc that are always comming down, so less power is being used each iteratiion we get on the cpu and other electronic components. That is what get as climbing. That's what happens when we shrink the processors each generation. There is less space in their gaps, so less time needed for propagation, less power and better skew rate since less voltage to climb and fall.
Evolution always happened like this, as you may notice, some advances are jsut not in the right time, but when we need it it's already mature enough to be used. Just look around and you'll see it happen everywhere in any dimensions you may think of.
Sorry for my english, I'm Portuguese.


RE: Odd
By zshift on 3/25/2009 10:04:14 AM , Rating: 2
very good point.
btw, your english was actually very well written, I would never have noticed that you weren't a native english speaker if you hadn't said so yourself.

still, even if it's not the most important factor in processor speed, it's still great know that some of the pieces we need to increase the processor speed are close to being fully developed and ready for commercial use.


RE: Odd
By TheBaker on 3/25/2009 4:18:25 AM , Rating: 5
"If they were able to make a baseball bat that wouldn't break even if it were used to hit a baseball 100 Miles, would major league players start hitting them that far? "

You're right in saying creating a high enough output frequency isn't the problem. Right now the limiting factor for CPU speed is heat. The heat is generated because of the amount of power required to filter out the signal noise mentioned when high multipliers are used. This material eliminates that problem.

So, to use your analogy, the baseball players can already hit a baseball 100 miles, but every time they try, the bat breaks. If you give them a better bat, then yes, they would start hitting them 100 miles.

This new graphene part is that bat.


RE: Odd
By MrPoletski on 3/25/2009 6:41:29 AM , Rating: 2
It's also very low power.

Were power consumption and the clock speed wall not something that stood in the way of the P4?

Consumed too much power because you need more to get the steady clock signal and hence produced too much heat.


Burst!
By Alphafox78 on 3/24/2009 12:17:07 PM , Rating: 1
Long live netburst!!!




RE: Burst!
By FITCamaro on 3/24/2009 12:22:47 PM , Rating: 5
Even at 1000 GHz it'll still suck balls.


RE: Burst!
By StevoLincolnite on 3/24/2009 12:58:16 PM , Rating: 3
Instead of heating the home it would end up causing global warming.


RE: Burst!
By FITCamaro on 3/24/2009 2:41:27 PM , Rating: 5
Why not. Everything else does.


RE: Burst!
By JoshuaBuss on 3/24/2009 4:56:05 PM , Rating: 2
*rimshot*


RE: Burst!
By GlassHouse69 on 3/25/2009 11:15:56 AM , Rating: 3
*money shot*


RE: Burst!
By MrPoletski on 3/27/2009 5:56:20 AM , Rating: 2
*HEADSHOT*


RE: Burst!
By johnsonx on 3/25/2009 8:17:49 AM , Rating: 3
It's Climate Change now you idiot, we don't say Global Warming any more. Crazy people got the weird idea that if there's no warming, then there's no Global Warming... so we had to call it something else that applies to everything.


Pentium 4 redux
By Jansen (blog) on 3/24/09, Rating: 0
RE: Pentium 4 redux
By Alphafox78 on 3/24/09, Rating: 0
RE: Pentium 4 redux
By FITCamaro on 3/24/2009 12:39:21 PM , Rating: 5
So we flog you first?


RE: Pentium 4 redux
By bjacobson on 3/24/2009 12:40:19 PM , Rating: 5
quote:
I posted about netburst waaaaay earlier!


This post brought to you buy an Athlon 64 processor!


RE: Pentium 4 redux
By DeSade on 3/30/2009 8:37:10 AM , Rating: 2
Not sure which is funnier, the overt or subliminal AMD advertising.


RE: Pentium 4 redux
By Storkme on 3/24/2009 12:37:38 PM , Rating: 4
The problem of high clocking isn't just to do with generating the clock, it's also to do with pipelining. If the clock signal is too high you won't leave enough time for signals to propagate through the gates before trying to fetch a result, and you'll get inconsistencies and instability.


RE: Pentium 4 redux
By mmntech on 3/24/2009 12:58:35 PM , Rating: 1
A PC that can cook a full English breakfast?

Could be interesting to see where this goes. I had figured that conventional binary computing was starting to reach its limits.


RE: Pentium 4 redux
By ViroMan on 3/24/2009 10:49:14 PM , Rating: 2
cook? Naw, thats when you start using liquid nitrogen to cool your CPU.


Speed of light
By bman on 3/24/2009 5:32:18 PM , Rating: 3
1,000 GHz would mean that a signal moving at the speed of light (300,000,000 m/s) would only propagate 0.3 mm every cycle. I doubt we will have 1000 GHz CPUs any time soon (if ever) since the perimeters of the smallest CPUs today are much bigger than 0.3mm.




RE: Speed of light
By resonantwa on 3/24/2009 8:32:34 PM , Rating: 3
It is possible. All that is needed is to synthesize excited bromide in an argon matrix. As soon as we apply a field, we couple to a state, graphene is then radiatively coupled to the ground state. I figure we could then extract at least 10^9 Hz


RE: Speed of light
By denobin on 3/24/2009 9:49:58 PM , Rating: 2
@resonantwa:
I can't believe no one else got the "Real Genius" reference (lest anyone thought you were serious.) Geeks unite!


RE: Speed of light
By Black69ta on 3/25/2009 1:36:24 AM , Rating: 2
I smell Popcorn!


RE: Speed of light
By johnsonx on 3/25/2009 8:12:44 AM , Rating: 2
yes, but 10^9Hz is only 1Ghz. We've been there quite some time, even without synthesizing excited bromide in an argon matrix.


RE: Speed of light
By trisct on 3/27/2009 12:28:55 PM , Rating: 2
That assumes that to do useful work a signal has to traverse the whole CPU. Pipeline stages are much, much smaller, on the order of several microns. 0.3 mm is a gigantic functional circuit at that level. A deeply pipelined CPU (like NetBurst was) could make use of high clocks, but more radical architecture enhancements would be needed to get to a terahertz. Then you would have the issue of keeping this beast fed.

Like Itanium, you would start being limited by how much fast cache you had around the CPU.


RE: Speed of light
By Justin Time on 3/31/2009 3:34:40 AM , Rating: 2
Electricity doesn't travel at the speed of light... closer to 75% of it... so the issue of signal propogation at 1000Ghz is even more significant.


1,000 GHz = 1THz
By sdoorex on 3/24/2009 2:30:58 PM , Rating: 5
Wouldn't it be easier to state Terahertz instead of saying 1,000 Gigahertz?

Sorry, I hope this didn't double post.




RE: 1,000 GHz = 1THz
By xRyanCat on 3/24/2009 6:39:49 PM , Rating: 3
Marketing and headlines. More numbers always looks better. (Unless you're selling something.)

1,000 looks better than 1. And a quick glance people may not know what a THz is, while even Dell shoppers will see 2.4 GHz when they buy their processor and can relate it to 1,000 GHz.


RE: 1,000 GHz = 1THz
By inperfectdarkness on 3/24/2009 7:01:50 PM , Rating: 2
my post was 2 hours earlier...but same essential idea:

i submit to being flogged 1st now.


RE: 1,000 GHz = 1THz
By TheBaker on 3/25/2009 4:20:36 AM , Rating: 2
RE: 1,000 GHz = 1THz
By rcc on 3/25/2009 12:00:07 PM , Rating: 2
Once it becomes common place and people are used to it, it'll be TeraHertz. Until then, it would just confuse the masses


Pentium 5
By gamerk2 on 3/24/2009 4:58:58 PM , Rating: 3
Now introducing the Pentium 5 processer with Netburst 2.0, with speeds up to 1 terrahertz! Only $9,999,999.98!




RE: Pentium 5
By CommodoreVic20 on 3/24/2009 5:02:16 PM , Rating: 4
Too bad the hard drive is still working at a slugs pace.


hellz yeah!
By inperfectdarkness on 3/24/2009 12:32:18 PM , Rating: 2
can't be much longer before we have 1Thz FSB.




RE: hellz yeah!
By Sahkuhnder on 3/24/2009 8:34:23 PM , Rating: 1
Even will all that speed does anyone want to guess if Windows will still take a for....ev....er to slowly boot up?


RE: hellz yeah!
By ViroMan on 3/24/2009 10:58:15 PM , Rating: 1
Microsoft will find more ways to put crap in there that HAS to load be default because its useful to someone even though its not likely that you will ever use it.

(Universal Plug & Play sucks up tons of resources and may 10% of people use it [BTW its not the same as Plug & Play])

Ohh how about drive index... ya thats useful, not unless you search your drive often, and then I have to wonder how you can remember how to use a computer if you can't remember where you leave files.


RE: hellz yeah!
By marsbound2024 on 3/25/2009 7:34:57 PM , Rating: 2
If I recall correctly, FSB has been dead since November more or less.


Exciting
By Chiisuchianu on 3/24/2009 12:27:57 PM , Rating: 2
This, plus SSDs massive storage potential, plug 64bit PC, plus all the other stuff makes it look like a big advance in PCs are coming soon. Now all we need is for crappy ISPs to do something about our slow as molasses internet.




RE: Exciting
By TheFace on 3/24/2009 4:37:03 PM , Rating: 1
There are always huge advances in PCs coming. Especially in hardware. What needs to happen is better software to take advantage of ALL the processing capabilities of a computer. Starting with the OSes.


RE: Exciting
By snarfbot on 3/24/2009 9:35:35 PM , Rating: 2
well whats nice about clock speed is that it doesnt need any software magic to use that extra speed, its perfectly linear scaling, barring any bottlenecks from other parts of the system.

even if this can only double the clockspeed possible after theyve extracted every last bit of performance from the technology it would still awesome, because as it stands were hitting a wall in terms of clockspeed, and making things smaller isnt really helping anymore.


Cryisis Jokes?
By choadenstein on 3/24/2009 3:36:33 PM , Rating: 3
Come on, it's been this long and not a single Crysis joke?




RE: Cryisis Jokes?
By choadenstein on 3/24/2009 3:37:07 PM , Rating: 2
err.. Crysis... Guess the joke is my spelling.


But..?
By Clauzii on 3/24/2009 1:17:40 PM , Rating: 4
Didn't IBM already make a 500GHz transistor 3 years ago?

http://www-03.ibm.com/press/us/en/pressrelease/198...




1,000 GHz = 1THz
By sdoorex on 3/24/2009 2:30:24 PM , Rating: 1
Wouldn't it be easier to state Terahertz instead of saying 100 Gigahertz?




RE: 1,000 GHz = 1THz
By 67STANG on 3/24/2009 6:34:19 PM , Rating: 3
No!

Now if you don't mind, I'm headed to the store to buy some 1,000GB Hard Drives.


Core 2 i7
By DjiSaSie on 3/25/2009 4:26:03 AM , Rating: 1
Implements this to Core i7 processor system and virtually we'll have a multicore processor featuring multi terrahertz speed = a home version of super computer
And we may name it Core 2 i7 ?




RE: Core 2 i7
By cfaalm on 3/25/2009 6:45:16 AM , Rating: 3
Phenom III


*=1/division
By Casual Observer on 3/24/2009 12:53:02 PM , Rating: 2
So we bring you fire! oops wrong age!
We bring you frequency multiplication (rather than division) FFT, hwd. Neat a look ahead cruncher.




That's amazing...
By swhibble on 3/24/2009 8:43:06 PM , Rating: 1
BUT CAN IT PLAY CRYSIS?????? </obligatory crysis reference>




RE: That's amazing...
By riku0116 on 3/24/09, Rating: 0
RE: That's amazing...
By gstrickler on 3/25/2009 1:59:39 PM , Rating: 1
Not without cryo-cooling the 500GHz GPU.


Oh, "multiplier"
By rzrshrp on 3/24/2009 12:21:46 PM , Rating: 1
I clicked on the link to see what multiplayer would require 1,000ghz...then I looked a litle closer.




"Intel is investing heavily (think gazillions of dollars and bazillions of engineering man hours) in resources to create an Intel host controllers spec in order to speed time to market of the USB 3.0 technology." -- Intel blogger Nick Knupffer

Related Articles













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki