Slashdot Mirror
Experimental Transistor Breaks 600 Gigahertz
neutron_p writes "The goal of a terahertz transistor for high-speed computing and communications applications could now be within reach. A new type of transistor structure, invented by scientists at the University of Illinois, has broken the 600 gigahertz speed barrier.
A new type of transistor - built from indium phosphide and indium gallium arsenide - is designed with a compositionally graded collector, base and emitter to reduce transit time and improve current density. With their pseudomorphic heterojunction bipolar transistor, the researchers have demonstrated a speed of 604 gigahertz - the fastest transistor operation to date."
What's the power usage on this thing? For one transistor it doesn't matter too much, but remember that todays chips have billions of transitors in them- Intel's Prescott core is rediculously power comsumpive right now. Even worse, over 100 watts of the power is lost to heat! So, what's the power and thermal design power of these things?
I suspect the power requirements for this bi-polar transistor would make it impossible to build something on the scale of a P4.
Bipolar eats power.
I think these transistors, if found to be manufacturable, will probably be used in communications not digital logic.
Raydude
If I found out who you were, I'd come bitch slap you.
First of all, this transistor was developed in the ECE dept, not CS.
As for CS, we're not droping in rankings. On the other hand, we shall be climbing very soon. We have the highest percentage (and number) of young faculty of any CS program in the world. I give us 2-4 years before you see the results of the rampant hiring over the past 2-3 years (15+ new faculty members), who are all pushing to get tenure over the next half-decade.
Finally, Mosaic, the original httpd and NCSA telnet were all "intented" at NCSA, not at CS department. However, they did hire tons of CS grad students on those project. Nonetheless, these 2-3 inventions are not the only things that make us prestigous, except maybe in the eyes of mainstream media and the lay man.
now supporting:
cmdrTaco for president '04
michael for oval office intern summer '05
IMO you are correct that Intel/Microchip/AMD aren't going to change their processes without a damned good reason. But PC parts and logic circuits aren't the only thing transistors go into.
How about the RF modulators/demodulators in all cell-phones, the RF amps in same, the special-purpose chips, regulators, detectors, buffers, amplifiers, etc that mfgrs still crank out by the butt-load, etc?
Personally, I'd really get off on an op-amp designed around these puppies! Imagine the gain-bandwidth product (eff-sub-Tee)!
Cloned foods give the statement "We had that last week!" a whole new meaning.
I just looked on Agilent's website and they don't seem to have any 600,000 MHz oscilloscopes for sale. I wonder how they tested this thing? A string of divider flipflops, perhaps?
Nexsan Technologies SATA RAID
Assuming it's all about network communications and storage. I'm guessing you've never done any graphical apps with your machine. Those eat CPU's for lunch. I can peg even the fastest consumer machines available now with some of the stuff I work with. Just depends on what you do.
My blog. Good stuff (when I remember to update it). Read it.
One other point: transistors on exotic processes like these are most suited for wireless and communications applications at this point; problems with power density, logic density, yields, and manufacturability have thus far kept them isolated to radio and interconnect modulation systems to date, rather than bulk logic. These problems are not necessarily insurmountable, but prevent its use in fast microprocessors for the foreseeable future.
Of course, to put in into perspective, most LED's are Gallium Arsenide as well. LED's are packaged, and high frequency (>10 GHz) chips are ussually not.
It's not wasting time, I'm educating myself.
I seem to recall from performance theory that the only true benchmark is how well a machine can perform your workload (where 'your' means the user / buyer / operator / etc.).
If your workload is to see how quickly you can load google's website, well, I just don't know what to tell you. Maybe take a screen shot, print it out, and paste it on the monitor? Voilà, execution of your workload in 0 seconds! Every time! Repeatable!
Sorry, I just don't see why this 'benchmark' is relevant. Funny, yes. Relevant, hmmmm.
In Soviet Russia, us are belong to all your base.
Every mixer circuit (needed for heterodying) that I have ever seen uses transistors. That is my point.
Actually diodes are more common than transistors in high frequency mixers, and fast diodes are less challenging to build than fast transistors, so it wouldn't suprise me if there were already diodes in existence that are faster than these new transistors.
No, because unlike software engineering electrical engineering has to do with physics. For instance, the engineering required to crank up PCI-Express from 2.5 Gb/s to 10 Gb/s basically requires a complete reengineering of the whole physical layer circuitry.
It doesn't have to do with the semiconductors so much as the physics of the wires, which really screw up signals at that rate thanks to frequency-dependent loss mechanisms.
Lacking <sarcasm> tags,
Bad for the planet? The planet doesn't care. We're just a thin scum on the surface.
Socialism: a lie told by totalitarians and believed by fools.
Most of your more run-of-the-mill silicon chips are doped with arsenic as well. Phosphorus and arsenic are probably the most common N-type dopants used in semiconductor manufacture.
seven two six five
seven four six one seven
two six four two e