Motorola solves 30-year problem by accident

[camaro]

I was watching the story on tech TV. That technology is capable of 70GHz down the road.

[pbharris]

be kinda neat to see what a powerPC can do with that. Oh boy am I glad Linux runs on those too !!!

[t048]

So, what does this mean to me? Am I going to be able to get a 70GHz system in the next year or two? 70GHz is impressive, but is it only a "laboratory" speed or a real world speed?

BTW, I hear the space shuttles still run 486 cpus http://www.sysopt.com/news/gws/cdata/smile.gif

[OuTpaTienT]

There's a very specific reason for the Space Shuttle running that type of technology...probably too long and involved to get into here, ask in the forums and we'll explain better.

[camaro]

That 70GHz will take some time. But expect to see the first chips in full production within a year. I guess it's all up to Intel and AMD to work it into their roadmaps. When this does happen, software will be so far behind the hardware, you might as well flip a coin as to which cpu to buy. Will video cards have this also? Nvidia core running at 10Ghz on a AMD running at 10Ghz. That would be something interesting. Talk about memory/AGP/PCI bottlenecks.

[This message has been edited by camaro (edited 09-06-2001).]

[TechJumper]

This just goes to show you how much we've been screwed by AMD and Intel http://www.sysopt.com/news/gws/cdata/wink.gif

[caddmannq]

Well, this mega-gigahertz stuff is all just pie-in-the-sky until we see some chips on the shelves. But that's not what interested me enough to post this story.

This was my interest:

Scientists would often have you (and, more importantly, those that fund their research) believe that science progresses in a relentless ploding scientific method. A infers B, B leads to C, C proves D, etc. True, this is how you must go about it; but this, like most of the great discoveries in history, was made while looking elsewhere.

In otherwords, it was made by accident!

[DanU]

Don't trade your AMD stock for Motorola stock just yet....

Gallium Arsenide makes for extremely fast n-channel transistors, but only mediocre p-channel transistors. Mate GaAs NMOS to GaAs PMOS to create CMOS logic and you end up with a circuit that's no better than Si CMOS.

The articles state that this will mostly affect communications chips, to which I completely agree. Ultra-fast GaAs NMOS transistors are used for the RF front-end of a cell-phone, pager, microwave amplifier, etc... but separate chips using Si are needed to handle the "baseband" (i.e. voice, data, video, etc). Now you can have everything in one chip. This will enable these devices to be much smaller and cheaper.

BTW, science and serendipity often go hand in hand. A good read is "Serendipity: Accidental Discoveries in Science", by Royston M. Roberts.

[Ranma]

DanU,

Processors never utilize the same number of n channel transistors as p channel. For one, p channel transistors are about three times the size of n channel transistors, making them a considerable burden. Instead, CPU fabrication designs have long included implementations of dynamic switch networks: charging the switch network to a high voltage for part of the clock cycle, then using a pull-down n channel network to bring down the voltage if necessary.

The development of p channel transistors did not parallel the inception of n channel transistors, yet processors and other devices were implemented using only nMOS technology.

Welcome to tomorrow-- the past.

[majik_mushi69]

I think we shouldn't be going for speed so much, But wider data buses.
personaly I would like a 1024bit processor @ 1 Ghz than a 64Bit Processor @ 16 Ghz.

[IW]

Often politicans and the American public both have a problem with funding for science that doesn't seem to have any use.

Serendipity is one of the main reasons that we shouldn't be so quick to complain and not fund things that seem to not offer some instant payoff.