8GHz (with the help of liquid nitrogen) not quick enough? Leave it to the folks at MIT to make sure your zaniest desires are well taken care of. As research forges ahead on
graphene, carbon nanotubes and buckyballs (remember those?), gurus at the university have discovered a breakthrough that could eventually lead to microchips that make existing silicon-based CPUs weep. In fact, the research could lead to practical systems in the 500 to 1,000 gigahertz range. The magic all ties back to advancements on a graphene chip known as a frequency multiplier, and while the nitty-gritty of all this is far too complicated for the layperson to grasp, all you really need to know is this: finally, you can rest assured that you'll one day own a chip capable of handling
Duke Nukem Forever.
[Via
InformationWeek]
posted Mar 26th 2009 8:03AM
What's the wattage? 12.1 jiggawatts?
with that you could power 10 time machines... what a waste...
Oh, the possibilities!
12.1??? Great Scott man, what's wrong with you, it's 1.21. Where the hell would you find 12.1 jigawatts? Some people...
I apologize to the Joystiq community for ruining what could have been an awesome post.
I sincerely hope that you can forgive my accidently placing a decimal in the wrong place.
A: Joystiq community??
B: The only thing that can produce that kind of electricity is 10 bolts of lightning!
... or Mr. Fusion
Joystiq? Wow, someone must have forgot their V8 this morning...
Nah, I've been awake all night. Couldn't sleep.
I hate V8.
No, V8's are for autoblog.
Grammer Nazi's are from Joystiq.
Apple trolls are engadget.
Isn't a V8 a car?
*grammar
COME GET ME! >:)
ImaYam = Joystiq Troll Bait.
Come get me! Im ARM'ed with flying New Gen Shuffles VIA my shoulder mounted, usb powered AIR rifle!
Approximately same power req or a fraction higher as current cpu's that's why the research into this material for cpu production is done. Silicon is nearing the end of it's curve.
Regardless the frequency/powerdraw/heatproduction ratio is far better then silicon.
Hot damn!
Duke Nukem forever, you say!
oh yeah!!
@Imayam
For some sad reason after your last comment I believe you live in a dungeon.
Im bored of these 'discoveries' that you hear about but then never materialise.
I'm not, please get out my internetz
I'm not bored of the discoveries; I'm bored of the lack of materialization.
Key word is "Eventually"
Eventually, this will be implemente into everyday retail, and but the time that happens (I estimate 10 years), it will seem no more significant than the introduction of the core i7's. It's the discoveries that are truly interesting - by the time they are materialized, and many are, much of the initial enthusiasm has simply decayed over time.
But will it play Doom?
A better question would be:
Will it blend?
those are both horrible questions
Will it be able to fabricate the fantasises of geeks around the world with its sheer processing power?
Will it be able to party?
will it play mario?
i mean will they replace human playing mario?
I wouldn't expect it in production earlier than 2025, but I would love to be proven wrong.
maybe it won't be so long :
"he thinks developing it to a stage that could become a commercial product may take a year of work, maximum two."
Bu...Bu.....BUT INTEL and AMD said that physics nature/dimension return prevent creating past 4k CPU and we gonna stick to the rest of our life with 3k-4k cpu (Since a decade We are stuck with High 3k Cpus~ they said multiply the cores is the way to go
Who is the lier?
It's liar, and it depends on your point of reference. I can only assume that when Intel and AMD said that, if they said it at all, it was within the context of silicon chips. These are chips made not of silicon but of graphene. The whole reason they are looking for new materials is that Intel and AMD would have been right, based on what we know of silicon. You can't just keep making everything smaller, at some point it can't get any smaller, or it isn't cost efficient to try. Graphene apparently allows us to go past that limit of silicon.
So if they said that, I wouldn't call them liars.
GRAMMAR NAZI ! j/k
Intel and AMD said what i wrote up many times over the years, i suck with Google search , I'm sure someone else can Google it and get the info related to this subject
I prefer to get 1 Core cpu with 10k speed over 8 cores 3k speed
My dual core 2 duo 3000 (OC to 4k And its stable) is way faster than my Quad Core 2.400 Cpu "OC to 3.200" (because most Games/applications is just starting to support dual cores (Almost mainstream now) only few Games/Programs "Like video editors etc" starting to support Quad Cores
Release a Dual Core 10.000 Speed And humanity wont need new CPU for the next 10 years
You!
Its a YO.....
DR House.
Those comments are made in the Current Material science being used for making semiconductors in Mass production.
u
Given that a fab is costing close to 3 billion to build retooling machines and or fabs will not be a trivial undertaking. (even the Low K transition costs millions of dollars to implement and that is a VERY VERY small part of making a chip).
Given that there are millions of Man Hours experience with Silicon development moving to new materials is also a non trivial task as Engineering will have to adjust to this new platform. and there are several years left in Silicon before we meet the Physical limits. (about 15 to 20 years before we hit this economic downturn).
Oh wow you have no idea what you're talking about do you? First of all, it's MHz and GHz. You have a dual core running at 4 GHz and a quad core at 3200 MHz. Anyway games are already limited by the dual cores and if you try running crossfire with the most powerful cards out there you will be limited by two cores. Most programs are starting to utilize all four cores and your dual core at 4 GHz just wont cut it anymore.
Also, why would you want a 10 GHz single core CPU over an octa core running at 3 GHz. Once they release the octa cores the software makers will code their programs to utilize all eight cores and you will have performance equivalent to 24 GHz(a bit less because efficiency is never 100% with multiple cores).
I hear you'll be able to run Duke Nukem Forever with Water-Shaders on Medium-High!
I wonder how many airports intel or amd will have to buy MIT to acquire this tech.
How many airports does MIT need?
A better question would be: "When did airports become currency?"
That was a weak Duke Nukem Forever reference
Your face, your ass, what's the difference?
Quite the opposite, my friend, very spot on! ;)
This article is incorrectly summarized. It's the *transistors* that can reach 1000Ghz, you cannot estimate wafer/chip frequencies by looking at single transistor on off speeds. IBM made transistors reach 500Ghz 3 years ago, that doesn't mean that we'll be seeing 500Ghz processors soon.
I agree on that statement.
Several things missing (btw there are Silicon transistors that are switching at 1 teraherz also
http://www2.electronicproducts.com/Terahertz_transistors_allow_continuation_of_Moores_Law-article-FEBOL2-FEB2002.aspx is one example. some of that technology is used it the current generation of chips.
there are such things a propagation delays which is a major limiting factor. also looking at the physical structure (hexagonal over cubic I can't see packing transistors as close (although carbon is a smaller molecule which will allow for smaller structures.
Now for the limiting factor of silicon circuits. There is no mention of heat on the switching transition. and what happens when you put Millions of these transistors 40 nanometers apart from each other. or even better 10 nanometers apart which will be around the limit that we can take silicon transistors to.
Well, I'm no physicist, but this quote from the original artile may go a little way to explaining why this technology has so much expected of it:
"Frequency multipliers are widely used in radio communications and other applications. But existing systems require multiple components, produce "noisy" signals that require filtering and consume large power, whereas the new graphene system has just a single transistor and produces, in a highly efficient manner, a clean output that needs no filtering."
http://web.mit.edu/newsoffice/2009/graphene-palacios-0319.html
I'm not sure if that gets around the transistor/chip difference, or the physical limitations of chip density, but to me it sounds like you would see minimal reduction in speed from the transistor to the chip technology.
Maybe it is jumping the gun, but the folks at MIT seem to think that "the new graphene technology could lead to practical systems in the 500 to 1,000 gigahertz range".
I suppose we'll just have to keep an eye the story to see how it develops
As long as they actually come out in our lifetime, seemingly unlike Duke Nukem...
The chip consists of one transistor and "could be commercialized in a year or two" right... I'm not holding my breath
Finally a CPU that runs Crysis at the highest settings...