P4 Prescott = slllooowww

[Goatman]

well, in comparison to AMD counterparts... lol

It looks like we have a new Williamette core.... poor poor intel fanboys...

Here are some Benchies.... discuss... what you guys think??

http://www.hardocp.com/article.html?art=NTgz

 

[Sazar]

the core is designed for raw speed... keep in mind as they ramp up the clock speed it WILL be fast...

clock for clock it is slower right now but @ the same time it's not exactly designed to be faster...

 

[Heeter]

It has always made me wonder why these companies diddle around with clock speeds nearly identical to the competition. Why doesn't Intel just a death blow to AMD and release a 4.5gig or 5gig CPU, or AMD do the same thing and watch the competition crumble? Well I know AMD is still on the 13micron build, but if I was Intel, I would release the 5gig, 1600mghzFSB right now and charge $1000 for it and watch AMD run and hide.


Heeter

 

[Sazar]

intel can't do it for a few reasons...

1 is technology.. they don't have the capacity to release a cheap and effective solution to hammer AMD like that.. just look @ how much heat the prescott core produces... do you really think they are capable of producing a room temperature running cpu @ those clocks? further what makes you think amd can't counter with a clock bumped a64 core ?

2 is costs... amd is surviving and now turning a profit because it produces higher performing/lower cost low/mid range products... durons are upto 50% faster than celerons and cost far less... which would you buy ?

amd can't do it for the main reason that its core is not designed to be ramped up in clock speed like intel;s...

prescott has 31 pipeline stages and high latencies and less work per clock cycle than northwood for one specific reason... CLOCK SPEED... it is DESIGNED to ramp up rapidly..

amd on the other hand has a product that does I spose now almost 2 times more work per clock cycle when considering prescott...

were it easy for top companies to produce a part to crush the other... they would have done it already m8

its not as easy as you think..

 

[Goatman]

But if they are having heat problems with it running at 3.4 think of what will happen when they try and ramp up the clock speed. I think it's just a stop gap measure for their next big release.

 

[Sazar]

the process will only improve goatman.. and so thermal issues will likely be solved..

they have to be for intel's prescott to be a success since its designed to ramp up to extremely high clocks...

 

[Sazar]

am surprised given the lacklustre performance of this processor more people have not chimed in

 

[LeeJend]

Intel shot themselves in the foot recently. They decided to stay with existing silicon on silicon process expecting to beat the heat issues with lower loss dielectrics and the smaller transistor geometries in 90 nm processes.

They FUBAR'd. The dielectrics weren't that much better and the leakage currents have gone up faster than junction capacitance has gone down with transistor size.

AMD and others are using newer silicon on insulator processes which get rid of the leakage (heat) problem. AMD has a chance to catch up to Intel on clock speeds and start getting market share back over the next 2 years while Intel admits defeat and switches processes.

The same thing happened back in the ****ty P3 days when Intel's weak architecture held them back while AMD raced ahead on throughput and clock speed.

Of course with Intel's massive capital reserves they can throw money and manpower at any problem and recover faster and better than AMD could even dream of so any AMD resurgence will always be short lived (but highly enjoyable).

 

[Sazar]

intel is using strained silicon m8... that is a step up from 'normal' silicon and a step in the right direction..

intel's prescott will likely show more muscle as its process is refined but for now amd rules the roost... tejas should change things a bit with its on-die memory controller

 

[Goatman]

But in the past die shrinks have always ended up running cooler, and more efficient. Core tweaks will only drop temps slightly.... I think Prescott is just a stop gap measure for Tejas.

 

[Sazar]

its not always cooler with the first batch of a new iteration...

recall when amd went to 0.13 micron with their 2200+?

you could barely oc' the sucker but subsequent batches allowed for a much better ability to oc the processor...

 

[Goatman]

I quote from Overclockers.com

The power consumption figures are terrible. From these, you have to assume at best a 130W 4GHz Prescott. Any overclocker adding just 10% extra voltage is looking at 160 watts. That's just for 4GHz. Mr. Freezes should have no problem exceeding 200 watts.

he compares what die shrinks have done to intel CPUs in the past, to what happened to Prescott.

read the rest of the article here

http://www.overclockers.com/articles948/

also here

http://www.overclockers.com/articles952/