What happened to that Computer Scientist's Law of CPU Speed?

[Nicotine]

Is it just me or do CPU speeds seem to have stagnated in the 3Ghz area...?

What's up with that?

[squall458]

dual core is the next thing, u mean moore's law (spelling?) ? its still increasing but thats not the focus anymore given that intel cant keep up with heat, and amd is beating them hands down while not using the increase clock speed trick. its all about efficiency now which makes me happy. my amd 2500+ mobile can smoke any comparable intel cpu.

[jerryhung]

Yes, it is Moore's Law
http://en.wikipedia.org/wiki/Moore's_law

Honestly I don't care, as long as throughput of CPU power is increasing, it's good enough. Who really needs 3, 4GHz anyway :)

[EchoAngel911]

moore's law is just about the growth rate on transistors , not the actual cpu speed

[brooksjames]

the next great advances are going to come with software advances not with just throwing more power into a cpu.

after 3ghz really whats an extra 500mhz or even more? i get the feeling we're not going to ever see that exponential growth in computing power (not saying clock speeds--- they're ineffective at measuring power anyway) ever again at the speed we saw it during the 90s and first 5 years of this century.

[ynchu]

Is it just me, or everyone else thinks so...?

Moore's "Law" should just be called an observation. Plus,
I thought AMD's processor design does not follow Moore's Law...

[B0000rt]

Is it just me or do CPU speeds seem to have stagnated in the 3Ghz area...?

What's up with that?

Clock speeds have little to do with speed anymore.. :razz:

[sumfunny]

I thought the law was that processors doubled speed every 18 months. Not true anymore because of the current limitations on cooling etc.

[Headhunter]

Moore's "Law" should just be called an observation. Plus,
I thought AMD's processor design does not follow Moore's Law...
Law sounds more authoritative; would you have remembered "Murphy's Observation"? Probably not. :)

[balou911]

Current CPU production technology has reached its limits.

IBM and others are working on commercial CPUs with 10GHz+ clock speeds.

Why does it matter anyways, with nearly 4ghz CPUs, DDR2 dual channel and PCI-express, the bottleneck is now your hdd.

[MameXP]

Guys, clock speed mean...... NO ****!

What matters is throughput thats how many instructions can be done in a specific time, also how fast can one instruction excuted? (higher CPU clock speed doesnt really matter here as it depends on also many variable such as pipeline,...etc)

So you said we're stuck? Hell no we still moving pretty smoothy with 64bit extension, hyper transport bus, dual core....

Also mobile 2500+ doesnt beat "any" pentium4 LOL. even at 2.7ghz . Do your homework . Its probably because you dont see a real "fast" pentium4.

[divx]

Guys, clock speed mean...... NO ****!

What matters is throughput thats how many instructions can be done in a specific time, also how fast can one instruction excuted? (higher CPU clock speed doesnt really matter here as it depends on also many variable such as pipeline,...etc)

So you said we're stuck? Hell no we still moving pretty smoothy with 64bit extension, hyper transport bus, dual core....

Also mobile 2500+ doesnt beat "any" pentium4 LOL. even at 2.7ghz . Do your homework . Its probably because you dont see a real "fast" pentium4.

he said any "compareable" p4, learn to read before you speak.

[morkys]

Factor in the limitations of materials science and manufacturing process and the law of diminishing returns relating to the scale of the PC technology and how they limit the possiblity for exponential growth. There's only so much faster you can make them. At 100 mhz, it was easy to up PC's to 200 mhz, and 400 mhz, and even 800 mhz, each time doubling the PC's speed. Once we get to, say 5 GHZ, do we have enough manufacturing capability to double a PC's CPU clock to 10 GHZ in a practical time span? Where will the limitations be reached? Certainly there will be a point where the manufacturing process will have to be changed, and we'll have to do something else to get things faster. I don't think we will see infinite increases in CPU clock speed.

[radeonboy]

Clock speeds have little to do with speed anymore.. :razz:

I concur. Remember the days when Intel said speed is everything? Well now they say speed isn't everything and all this time they have been pushing their hot P4's with higher speed "numbers". AMD on the other hand haven't been playing with high numbers but with high effiency.

The problem is now regular joes think the higher the MHz, the better. And Intel has gotten a huge marketing budget which means they have taken over everyones mind to say Intel when buying a PC.

I still go with AMD though; the underdog :D

[DaFonz]

Moore's law said that the transistor count would double every 18 months...

NOT SPEED.

So far, it's proven true.

[crimsona]

I mean, just look at the number of transistors on a graphics processing unit... Moore's Law has nothing to do with clockspeed anyway

[Gee]

As someone else mentioned. The bottle neck is the hard drive and other components.

But lets get the software to catch up first. Windows XP is slow. Use to be the software required faster hardware. It is now reversed, the hardware is so far ahead of the software.

I am still waiting for the day we load or entire OS in RAM and just access our files from hard disk.