500Ghz?!?!?!

[DaGuv]

Just noticed this over on theregister.com....

http://www.reghardware.co.uk/2006/06/20/ibm_overclocking_feat/

-268.5°C :O

 

[maverik-sg1]

Impressive............. but I bet they can't put a fruit pastel in their mouth without chewing it

 

[boardy]

Impressive............. but I bet they can't put a fruit pastel in their mouth without chewing it

I CAN!!! My mate challenged me to it the other day when I was hammered and I did it lol.

I saw the article earlier, was gonna post but cudnt b bothered as I thought every1 here reads the register anyway so will have seen it already lol.

Boardy

 

[DaGuv]

lol, you have more will power than me... i tried, and failed lol 8)

 

[Animal]

ok thats good but im more impressed at this...

chip running at 350GHz at room temperature

 

[K404]

I`m not sure if this will be a full-blown CPU, or just a simple clock oscillator. When I see Intel/AMD stuff at these speeds, I`ll believe it BUT (here comes the circuit and semicon theory)

lets assume a signal path is 1cm long

if the signal travels at the speed of light 3x10^8 m/s it will take:

3.33x10^-11 seconds to travel the path.

if a CPU has a clock speed of 350GHz...the pulse period is 2.85x10^-12

Thats less than the time it takes for the signal to travel 1cm.

If all the sections of the CPU are clocked simultaneously...the CPU will fall over "instantly"

For this to stand a chance, the sctions of the CPU will need to be asynchronous.

More theory to come, just ask! I have a degree in this now

 

[maverik-sg1]

I`m not sure if this will be a full-blown CPU, or just a simple clock oscillator. When I see Intel/AMD stuff at these speeds, I`ll believe it BUT (here comes the circuit and semicon theory)

lets assume a signal path is 1cm long

if the signal travels at the speed of light 3x10^8 m/s it will take:

3.33x10^-11 seconds to travel the path.

if a CPU has a clock speed of 350GHz...the pulse period is 2.85x10^-12

Thats less than the time it takes for the signal to travel 1cm.

If all the sections of the CPU are clocked simultaneously...the CPU will fall over "instantly"

For this to stand a chance, the sctions of the CPU will need to be asynchronous.

More theory to come, just ask! I have a degree in this now

Exactly my point

 

[nick25]

They could have at least run SuperPi 8)

350Ghz at ambient :?

 

[DonS1mpson]

I read somewhere that this wouldn’t be a chip, but a transistor of some sorts?



It proves us one thing though – The MHz war isn’t quite over yet