The Future of Computer Technology

aphexcoil

aphexcoil

These are exciting times in computer technology. Recently, Intel, IBM and AMD have demonstrated transistor technology that is on the nanometer scale (less than .1 microns). However, with the decreasing size of each transistor and more of them on each chip, the power requirements are becoming such that, at the present rate, chips 5-10 years from now will put out more heat than a thermo-nuclear reactor.

Intel is approaching this problem using a new design for transistors.

Intel's Press Release
http://www.intel.com/pressroom/archive/photos/20011126tech.htm


Intel has already developed the world's smallest and fastest CMOS transistors, including a 15 nanometer transistor, which will enable chips with up to one billion transistors by the second half of this decade. However, as hundreds of millions, and even billions of smaller and faster transistors get packed on to a single piece of silicon the size of a thumbnail, power consumption and the amount of heat generated in the processor core becomes a significant technical challenge. Using existing methods of semiconductor design would eventually lead to chips that are simply too hot for desktop computers and servers. These limitations could even prevent new chip designs from being implemented in smaller computers like mobile PC's and handheld devices.
More...

However, there are physical constants that cannot be breached. At these small levels, the transistors literally switch single electrons -- so they cannot act as amplifiers. The limit for this technology will fall between 5-9 nanometers, and then that's it -- we simply cannot build transistors any smaller than this. This will get us up to around 2020 with top speeds around 50-100 gigahertz.

Ahhh, but don't fret just yet. Once we get to the atomic level, really weird things begin to take place -- this is within the realm of quantum mechanics. At this level, electrons simply "sometimes" just tunnel right through the gates and some times just to something else. This is the nature of quantum mechanics.

Once we reach the limits of the "classical computer," quantum technology will already have advanced to a level enabling quantum computations. The really amazing thing about quantum computers is their ability to tap into alternative universes to use virtual computers among "billions" of other universes (Hilbert's Space) to make calculations using something called Qubits. Qubits can be both on and off at the same time!

Advanced Resource: http://www.informatics.bangor.ac.uk/~schmuel/comp/comp.html

Layman's Resource:
http://www.cs.caltech.edu/~westside/quantum-intro.html

These computers will be millions of times faster than today's classical super-computers.


In a traditional computer, information is encoded in a series of bits, and these bits are manipulated via Boolean logic gates arranged in succession to produce an end result. Similarly, a quantum computer manipulates qubits by executing a series of quantum gates, each a unitary transformation acting on a single qubit or pair of qubits. In applying these gates in succession, a quantum computer can perform a complicated unitary transformation to a set of qubits in some initial state. The qubits can then be measured, with this measurement serving as the final computational result. This similarity in calculation between a classical and quantum computer affords that in theory, a classical computer can accurately simulate a quantum computer. In other words, a classical computer would be able to do anything a quantum computer can. So why bother with quantum computers? Although a classical computer can theoretically simulate a quantum computer, it is incredibly inefficient, so much so that a classical computer is effectively incapable of performing many tasks that a quantum computer could perform with ease. The simulation of a quantum computer on a classical one is a computationally hard problem because the correlations among quantum bits are qualitatively different from correlations among classical bits, as first explained by John Bell. Take for example a system of only a few hundred qubits, this exists in a Hilbert space of dimension ~1090 that in simulation would require a classical computer to work with exponentially large matrices (to perform calculations on each individual state, which is also represented as a matrix), meaning it would take an exponentially longer time than even a primitive quantum computer.
More...


Today quantum computers have already demonstrated the ability to instantly find the factors of small numbers (due to their limited Qubits as of yet). However, this presents a large problem for cryptography since that technology depends on the simple fact that it is easy to confirm a number as a factor for a very large number but extremely CPU intensive to devise one from scratch. Quantum computers can reach into how many other universes are necessary to come up with EVERY factor instantly.

 
Quantum computing is still in its infancy. It has great potential and for some of the grand challenge problems in computing and physics it has unique application.

However one of the things to contemplate is this: Computing and storage is getting cheaper every year - we still have not really deviated from a moore's law curve. What will we do with all of the existing and future capacity ? IMHO the real issue for science to consider is that the raw computing and storage power per unit dollar that is now available is outstripping application need. Yes, there are some problems that inherently can scale to use massive computing arrays but by and large what is needed are new approaches to many of the outstanding existing problems. Much of science today is merely adapting exisiting probelms to ever larger computing resources rather than concentrating on approaching disciplines from fresh points of view.

The availability of cheap computing and storage is helping many disciplines to advance in ways that were not anticipated a few years ago but we are sorely in need of new approaches to harness just the existing computing power available.
 
Quote from CalTrader:

the real issue for science to consider is that the raw computing and storage power per unit dollar that is now available is outstripping application need.... we are sorely in need of new approaches to harness just the existing computing power available.
More...
I couldn't agree more. I've been telling people this for quite some time now. Thanks.
 
Yes. Trading with a 500GHz computer won't give you more profits than with a 500MHz one. The computing power of the current hardware is more than enough for the traders. But the Tech companys always want to sell us somethings new so they need to make fancy stuffs.
 
Quote from hermit_trader:

Yes. Trading with a 500GHz computer won't give you more profits than with a 500MHz one. The computing power of the current hardware is more than enough for the traders. But the Tech companys always want to sell us somethings new so they need to make fancy stuffs.
More...

Ok, but they can't sell it to you if you aren't buying. It's all mostly ego driven for the massess now. There is more computing capacity in the base levels of Pentium II hardware combined with maximum storage of ram and drive hardware than many folks will EVER need. It tickles me when I hear/see folks looking for 1.1GHz, 1.3GHz, 2.8GHZ, and now even 3.0GHz+ machines to do basic math. Sure it seems complex to you, but the computer resource meters never get to stay pegged to the maximum readings for most.

Two weeks ago, I took a few of one of my client's 500MHz PII machines out of his office. Put them in new cases, bumped all the RAM up to over 1 Gig for each machine. Added some great video cards and faster spinning hard drives. He swears those new machines are really great. I hated to break his heart with the news that he really didn't have faster processors. The sad piece is he called me today to say he wanted faster processors. For word processing and Internet duties? Go figure! :)
 
It sounds like some of you think the only thing computers are good for is trading :) . Remember when one of the really huge computer companies way back in the 50's I think it was, stated unequivocally that the whole world, at most, might need a dozen or so computers. It is only the limitations of our experience that create an incapability to imagine uses for computing power many millions of times faster than current performance. For those interested in learning more, get Kurzweil's book, "The Age of Spirtual Machines". It'll blow your mind. It's science fiction turned real. He poses some very interesting arguments for what it would mean to have virtually limitless computing power.
 
You must log in or register to reply here.
Back
Top