Is Intel Releasing a 64-Bit Processor with the Prescot Next Year?

[Krazie]

When will intel come out with a 64 bit processor? I am excited about the prescot, but will it be 32 or 64 bit? And does anybody know if the prescot will be out this time next year? And does anyone know what the clockspeed of processor will most likely be this time next year?

 

[dduardo]

The good news: Intel already has a 64-bit processor. It is called the Itanium
The bad news: Intel is scraping the P4 (I think this includes the prescott)
The good news: Intel is replacing the P4 with a dual core chip.
The ugly news: Don't expect this chip for a while.

 

[Foo Fighter]

I'm upset that processor speeds have not greatly increased over the past year. A year and a half ago, Pentium 4 3.0 ghz chips came out. Now it's 3.4 ghz. This is the slowest level of increase since as long as I can remember computers have existed. Although other than processor speeds, other things have progressed. Like the Front Side Bus increasing to 800 mhz and dual channel RAM coming out. The processor speeds I believe is suppose to start increasing again though. With 10 ghz processors expected within a couple years.

 

[dduardo]

I don't know if there is going to be any huge speed increases coming anytime soon. Heat and power consumption are really causing lots of problems. The smaller the die size, the more they have to ramp up the current due to leakage. This is why intel has decided to go with their Pentium M technology which focuses on less power consumption.

 

[Guybrush Threepwood]

this is for all the MHz maniacs out there:
let me tell you a simple fact about CMOS technology. Current consumption (that means power too) is proportional with the frequency squared. So before asking for XXX GHz processor you'd better think about how are you going to cool them?

 

[theamazingTWOeyedman]

yes, if intel put out a 10GHZ chip with their current manufacturing tech, it would a fan/heatsink the size of your case to cool it, plus you would have to shout over the noise... But hey you'd have the latest right? :uhh: :tongue2:

 

[hitssquad]

Voltage, current and clock speed

Current consumption (that means power too)

No, it doesn't. Power is voltage * current (or, more generally, force * rate of application of that force).

is proportional with the frequency squared.

No. "Power functions are proportional with the frequency and with the square of the voltage [3], [4]." The reason power increases as the square of voltage increases (within a given system) is that current in a circuit goes up as the voltage goes up. E.g., doubling the voltage also doubles the current, and both of these effects together add up to a quadrupling of power consumption. However, simply doubling a chip's clock speed will merely double the electrical current feeding it, and hence merely double the power consumption. Battling this, CPU's have gradually evolved to operate on ever-lower voltages. The power-draw is still rising, but not as much as it would without concommitant decreases in voltage.

So before asking for XXX GHz processor you'd better think about how are you going to cool them?

It seems that computers will soon begin shipping with built-in water-cooling systems. Compressor-based cooling solutions may lie in the future for commodity computer systems, as well.

 

[hitssquad]

The Cooligy solution to CPU heat

yes, if intel put out a 10GHZ chip with their current manufacturing tech, it would a fan/heatsink the size of your case to cool it, plus you would have to shout over the noise...

My small air conditioner removes 1500 watts of heat continuously, without making so much noise that you can't easily talk over it. As far as PC cooling today (or in the near future), Cooligy claims to have in development a rather quiet and extraordinarily effective liquid cooling solution for CPU's.

 

[russ_watters]

I don't know if there is going to be any huge speed increases coming anytime soon. Heat and power consumption are really causing lots of problems. The smaller the die size, the more they have to ramp up the current due to leakage.

This is the end of Moore's law.

My small air conditioner removes 1500 watts of heat continuously, without making so much noise that you can't easily talk over it. As far as PC cooling today (or in the near future), Cooligy claims to have in development a rather quiet and extraordinarily effective liquid cooling solution for CPU's.

When you have one computer in a room, noise might not be a big issue, but when you have 50, its a deal-breaker. Also, a/c units and even just water cooling add significant cost and complexity to a computer. And then of course, there is heat rejection: what do you do with all of that heat when it leaves the case (after being multiplied by 3 by a portable A/C unit)? Pretty soon, they'll have to circulate chilled water to every cubicle in an office for cooling the computers.

 

[Njorl]

So what is everyone doing that requires so much speed? Most of the slowdowns I experience are with the internet and sometimes even a local area network. The only local slowdowns I have are with extensive 2d designs, and that is because of the video card. I don't think I've ever noticed my 2GHz processor being faster than my 1.2Ghz one.

When you sit in a traffic jam every day, you're more excited about a new highway than a new sports car.

Njorl

 

[hitssquad]

g factor of computational power

I don't think I've ever noticed my 2GHz processor being faster than my 1.2Ghz one.

Clock speed is not the same thing as a general factor of computational power. Which two CPU's are you using?

 

[russ_watters]

So what is everyone doing that requires so much speed?

I'm currently converting my parents' old vhs tapes to dvd...

 

[BoulderHead]

Russ,
What codecs do you transcode with ?

 

[Guybrush Threepwood]

Battling this, CPU's have gradually evolved to operate on ever-lower voltages. The power-draw is still rising, but not as much as it would without concommitant decreases in voltage.

The lowest suplly voltage for current digital circuits is around 1.8V (my motherboard monitor shows 1.5V). Do you think it can go lower? Keep in mind that you still need to differentiate between 0 and 1 logic levels. Also lower voltage for the logic levels means more sensitivity to noise...