AMD 'clocks' FX processor at 8.429-GHz

9/13/2011 10:26 AM EDT

This is cool :-)

9/13/2011 11:56 AM EDT

That is impressive. Did they use silicon or Ge? Impressive feat. Yet, is speed the key factor now? MPU should look for optimality and not just speed.

9/13/2011 12:19 PM EDT

The implication of the AMD press release is that this is a plain vanilla CMOS silicon processor....as it is part of Bulldozer introduction it is probably in 32-nm SOI HKMG and would probably be normally specified at 1 to 1.5-GHz.

I understand that "overclockers" do their thing by bathing the processor in liquid nitrogen!!!

9/13/2011 5:14 PM EDT

I think this is a publicity stunt on part of AMD...I doubt we will see a commercial part operating at that frequency...if that operation requires 77K temperature then the real value of that record is pretty doubtful...you get more gain by running multicores these days rather then speeding up the clock...Kris

9/13/2011 6:08 PM EDT

Publicity stunt or not, it's pretty impressive to achieve almost 8.5 GHz on a CMOS microprocessor, operating in a real system, even though they used liquid nitrogen cooling to do it.

They also got very impressive overclocking numbers with consumer-friendly cooling systems, as you can see in the YouTube video of this event.

With phase-change cooling (refrigeration to -40 C), which is a real consumer system option, overclocking to 6 GHz or so is practical. Gamers will be thrilled!

9/13/2011 8:30 PM EDT

BS. Who cares? Another science project...Ship me an air cooled 4GHz processor and I'll be impressed.

9/13/2011 9:02 PM EDT

Very nice effort to see/show what can be done. I would like to have seen the setup..I would imagine that they had some serious cooling. I wonder what the DRAM memory speeds were and how much they had to slow down the external memory accesses? If they can get half that speed in a normally cooled machine then I would be very impressed.

9/14/2011 12:44 AM EDT

"Overclocking, for those that have a difficult time obeying Moore's Law." ;)

9/14/2011 2:43 AM EDT

Would have been nice to know how much power they were drawing at that frequency, as well as how much they jacked the core voltage

9/14/2011 4:32 AM EDT

@elPresidente

Yes it would. That is exactly the sort of information AMD doesn't give out. I will ask though.

9/14/2011 7:19 AM EDT

I feel the need for speed! Speed does count - multicore hogwash - that came out because they had difficulty getting the accelator pedal further down. Run those 8 cores at 1 Mhz or even 100 cores at 1 Mhz - then compare that to a 3 gigaherz machine with one core and tell me about speed doesn't count. It's all horsepower.

9/14/2011 9:12 AM EDT

How are they cooling down the system?

9/14/2011 11:04 AM EDT

@elPresidents

AMD came back with this: "The CPU voltage was 2.0125 [volts], the temperature was below -230C, and the power consumption was not measured."

-230 deg C is below the freezing point of liquid nitrogen as I remember so it is amazing that the packaging, interconnect and low-k dielectric materials don't just shatter at some point.

I guess you take the temperature down as you take the clock frequency up, and very slowly.

9/14/2011 11:25 AM EDT

thank you Peter, good data...I love the point about "power consumption was not measured", really?, AMD went to all these efforts of clocking 8GHz and didn't bother to measure power, hard to believe...Kris

9/15/2011 2:04 AM EDT

Yeah, I agree the 8.4GHz is impressive even with all that liquid N2/He cooling. It goes to demonstrate that the CPU maintained functional integrity even at those very high speeds without so much as a glitch. Very good, however the cooling requirements make it almost a demo product. If they can get it to run at half the speed with just air cooling then they got a winner on their hands.

9/15/2011 10:43 AM EDT

Every CMOS circuit operates significantly faster when cooled below 77K...what happens if you take any other processor and run it at such temperature? Kris

9/16/2011 8:07 PM EDT

thank you Frank, this is the link:
http://www.youtube.com/watch?v=UKN4VMOenNM

9/22/2011 12:11 AM EDT

Aren't the transistor characteristics very different at such low temperature? Don't designers have fast corners where chip works significantly faster? Sounds like a publicity stunt.

9/22/2011 12:20 AM EDT

I can just see them spraying Circuit Cooler on it. ;)

9/24/2011 1:31 PM EDT

Of course it's a publicity stunt. Nice to read about such feat though.

9/27/2011 9:04 AM EDT

GOOGLE and read people! It is a 3.6GHZ, 8 core processor, not 1 - 1.5GHZ core. Yes it was a stunt (probably by overclockers as much as AMD) and yes they did use liquid helium, but it still worked.

AMD state that by air/water cooling the chip should be good for 5GHz.

Multicores are one way to speed, but with added software complexity and/or plain algorithmic limitations, GHz is often king.

9/29/2011 10:04 AM EDT

Intel i7 2600K is routinely being overclocked by end users like you and me to 5 GHz. And its almost easy to get 4.5 GHz.