Why have cpu's been stuck in the 3 ghz range for years?

[biggiesmallz]

Why have cpu's been stuck in the 3 ghz for years? For awhile speeds were increasing quickly now it has seemed to come to a halt.
Any speculations?

 

[Bobthelost]

1) Heat, increasing the clock speed increases the heat, ie P4.

2) Physical contraints of silicone, there is a limit to how fast you can make it turn on/off.

 

[biggiesmallz]

Originally posted by: Bobthelost
1) Heat, increasing the clock speed increases the heat, ie P4.

2) Physical contraints of silicone, there is a limit to how fast you can make it turn on/off.

So how many years will it take to overcome these hurdles?

 

[Bobthelost]

Never will. They're now bottlenecks that have to either be sidestepped (ie dual core) or gradually worked on incrementally with better processes.

 

[dguy6789]

Understand that processor frequency is not what we increase anymore to make processors faster. The architectures change and allow for more work to get done at a given clock speed than any previous processor. A perfect example is the Athlon 64. At 2.6Ghz, it is faster than a 4Ghz Pentium 4. This goes to show you that frequency is not everything. We stopped increasing clock frequency because we have reached the limit or are very near it of what current manufacturing tech allows. Instead of increasing frequency and heat output, AMD and Intel are now making processors that are wider rather than deeper.

 

[Keysplayr]

We can all probably place a safe bet that most, if not all, semiconductor companies (AMD, Intel, IBM, Motorola, etc. etc.) are working to develop semiconductors using materials other than silicon. I have read articles here and there about various "breakthroughs" from certain engineering labs that claim they manufactured synthetic material that has properties vastly superior to silicon. How soon will we see this stuff? My guess is the military sees it now, and will filter down to us boneheads eventually (5 to 10 years maybe). IMHO.

 

[purpleparalyzer]

There was a thread floating around about cpu's using magnetic islands in the future. I would link it but i dont have time right now- anyways I think that would greatly if not eridicate most of the heat but i dunno what the cons are.

 

[t3h l337 n3wb]

Right now, the race for CPU performance is shifting away from raw speed and towards multiple cores. Later, when they get rid of the heat barrier, speeds will increase again.

 

[Maximilian]

Yes, long live the GHZ race!

 

[Cooler]

Intel could have contunied Netburst but to reach it goal speeds water cooling and phase change would have to be the norm. Most people dont want to speed and exta $200 + to run a cpu at stock speed.

 

[KBTuning]

i wonder what ever happend to IBM's Stretched Silicon that they were working on back when AMD was at 1GHz and IBM was only at 700???

i think Intel worked with them w/ it but im not really sure.