prescott overclocking

[CyberFed]

Alright folks I've been out of the game for awhile but I'm getting ready to build a new system, I've been going back and forth between intel and AMD 64 chips. I was curious how well do the prescott cpus overclock? I was kinda leaning towards a 3.2ghz prescott on a IC7-max3 mobo. Is this a good overclocking combo I'm looking at hitting at least 3.5ghz (higher if possible of course)

Thanks for your help in advance

 

[Lancelot]

Yes, I think the IC7-MAX3 is about the only board that's really 'current-prescott-ready' so to speak. I just switched mainboards from Asus P4P800 to the MAX3 and all the heat around the CPU-socket is gone now. The OTES PWM-cooling system is really doing a great job. It also allows upto 3.2Vdimm so I'm now looking at high 1:1 overclocks! It's really a well-built board so I say go for it as well...

 

[deathstar13]

i have a differant view point and ideas than others,so with that said heres my opinions.

dont buy a prescott unless you have above average cooling,preferabley h20 or better for you to actually see its ocing capabilty.i can get my 2.8E doing 3.5ghz just a tad over default vcore,but temps get crazy even with above average air cooling and high cmf 90mm fan on max. 53c idle and 64c load.

when i bought my mobo my list ran like this.

1-p4c800e-d,i ruled out due to high costs.
2.epox p4ca3+ i bought as had everything the asus did plus 3.3vdimm and $50 less
3.ai7 -ruled out do to not being able to handle prescotts heat generation from the excess current it uses.

im not a big max3 fan as its said to have voltage regulation fluctuations and issues over 250fsb 1:1
i see the cooling thing as a gimik mostly as it does work but for hardcore use it still needs ramsinks on the mosfets.

the asus needs a vdimm mod for over 2.8vdimm i think.

i doubt anyone here or else where will agree with me on my decision to go with the epox.its mid priced,has excelent voltage options,even tho is said to top out vdimm wise 3v even tho 3.3v is an option.i havent tested my new board yet.
its dead stable has anything and everything wanted feature wise.
even has ide raid on board.
also i can attest to doing minumum 260fsb so far.most do more in the 280-290fsb range.

i do like the intel choice over amd atm,as thats what i did.
things do look better for them now they have the nf3-250 mobos with locks on them.but its funny but intel is cheaper.and i like the memory bandwith it gives over the amd offering.

 

[Lancelot]

"im not a big max3 fan as its said to have voltage regulation fluctuations and issues over 250fsb 1:1
i see the cooling thing as a gimik mostly as it does work but for hardcore use it still needs ramsinks on the mosfets."

It DOES have ramsinks on the mosfets! Four large ones actually that can be seen thru the OTES cover!!!

 

[deathstar13]

"im not a big max3 fan as its said to have voltage regulation fluctuations and issues over 250fsb 1:1
i see the cooling thing as a gimik mostly as it does work but for hardcore use it still needs ramsinks on the mosfets."

It DOES have ramsinks on the mosfets! Four large ones actually that can be seen thru the OTES cover!!!

wow i missed that one big time after looking again at some more pics.

but still it has the voltage issues along with the 250fsb limited issue. ramsinks were only a small issue these are rather large imo.

its a good mobo dont get me wrong,just wasnt for me.

 

[Lancelot]

Ahh I didn't know about a 250FSB issue, but my chip doesn't wanna do 250FSB anyways, didn't do it on my P4P800 and doesn't do it on the IC7-MAX3. Vcore is a tad lower (while keeping Prime95+SETI stability) on the MAX3 though but that may be due to the fact it doesn't suffer from V-droop as much as the Asus boards do. I wanted the MAX3 primarily to deal with all the heat Prescott generates and I must say it's doing an awesome job at that. My main rig is on 24/7 so I was really worried about all that heat in the long run with summer coming up and everything! It's functioning as a SETI cruncher and fileserver and is sometimes used for gaming. Other than that I do most of my 'regular' stuff on a P4-M laptop...

 

[CyberFed]

thanks for the quick replies, i think im going to shy away from the prescott. I'm thinking a regular 3.2ghz on a p4c800e-d, would 3.6-3.7ghz be a safe o/c estimate with good ram?

 

[acesea]

hey guys, ive been looking around but havent found much in regards to the performance of a p4e (prescott) vs. a p4c. i understand the prescotts run hotter, but if heat wasnt an issue, say phase change was used, would a prescott oced just like a p4c perform better than the p4c?
also, what can be expected of the new prescotts using the lga on the new motherboards?
the bottom line is, i dont know whether to grab todays p4c and oc them crazy, or wait for the newer prescotts on the newer boards.

 

[deathstar13]

hey guys, ive been looking around but havent found much in regards to the performance of a p4e (prescott) vs. a p4c. i understand the prescotts run hotter, but if heat wasnt an issue, say phase change was used, would a prescott oced just like a p4c perform better than the p4c?
also, what can be expected of the new prescotts using the lga on the new motherboards?
the bottom line is, i dont know whether to grab todays p4c and oc them crazy, or wait for the newer prescotts on the newer boards.

ive not owned a p4c yet.
but i do feel a prescott on phase change will beat a p4c in ocing.let me clear another thing up also.once you even overcome the heat issue expect to run into heat issues with the mobo.
my ramsinks on all my mosfets are very hot at times now just at 1.4vcore and 3.5ghz on air.this is due to the prescotts useage of higher amperage.

but dont stress to much over that.as every wall has another wall behind it.i just try and look ahead and be prepared so i dont get frustrated,after buying a prometia and then this happens.its the law of ocing.

the 875 chipset mobos clearley werent designed to run the precotts and was even rumored forever before the prescotts release that no 865/875 would ever support a prescott.now i know why.

the new boards should should have many advantages over the 875 line.

 

[(HR)Titan]

but what can we do
limit is 875/876 chipset

and i am suprised what mobo manufactures waiting with crating some ''prescot ocing chipset''

 

[GST-97]

I recently switched from a 2.8E to a 3.0C.
I have pretty decent watercooling, and the 2.8E at 1.42V, 3.5 Ghz would run around 40-43*C idle depending on ambient temp, this 3.0C runs 28-31*C idle at 3.6 Ghz 1.55V

The [email protected] is WAY WAY faster than my prescott ever was. In fact, the only benchmark I have run that shows anything slower, is RAM bandwidth... which is odd to me.

I just hated seeing 40C+ temps, even though it wasn't having a problem running at those temps, now seeing 33-34C full load temps seems a lot better to me

I would recommend you go with a 3.0C or 3.2C northwood at the current time.
As for the comment about the AI7, my AI7 handled the prescott just fine. I never had any instabilities or outrageous board temperatures (PWM topped out around 52C)... of course with this northwood it tops out around 38-40C

just my $.02

 

[batboy]

IThe [email protected] is WAY WAY faster than my prescott ever was. In fact, the only benchmark I have run that shows anything slower, is RAM bandwidth... which is odd to me.

As for the comment about the AI7, my AI7 handled the prescott just fine. I never had any instabilities or outrageous board temperatures (PWM topped out around 52C)... of course with this northwood it tops out around 38-40C

Well, I think 52 degrees PWM is rather outrageous. My old Prescott ran the PWM about 10 degrees cooler on my IC7 than compared to my AI7. That's a significant difference.

I compared a 2.8C to a 2,8E and found the Northwood was faster at default speed. At 3.4 gig the Prescott was catching up, although the Northwood still edged it out in most benchmarks. The higher you O/C the Prescott, the more it starts pulling away from the Northwood. Check out this link for a summary of my findings.

http://www.ocforums.com/showthread.php?s=8c054a1c5ddf290bcae07b333b4a450&threadid=289117

 

[GST-97]

Batboy: yes, I will agree that 52C is rather high, but I have seen people claiming 80*C+ PWM temperatures

I had my 2.8E at 3.5, and every benchmark I have run so far, this 3.0C @ 3.6 completely blows it away (except for RAM bandwidth in sandra/aida32).

Here are a comparison of the 2 processors in my system.
The [email protected]:

the [email protected]:

As you can see, my 2.8E at 3.5 was nowwhere close to this 3.0c @ 3.6

So whether I had something set up wrong with the prescott or whatever, this one does everything faster, and cooler

We also had figured out you were using an old version of sandra, so the numbers in that thread for your prescott is not comparable to the newer versions results.

-Chris

 

[batboy]

Your comparison is not exactly scientifically valid since you are not matching up same FSB or CPU clock speeds. Besides, Sandra is only one benchmark and certainly not the best either. Of course a 3.6 gig will beat a 3.5 gig. If you ran both at 3.5 gig, then you will see they are close to the same in performance... each winning some and a near dead tie in others.

I understand that a Northwood runs cooler and if you can run your new Northwood faster and cooler than the old Prescott, then you definitely made the right choice. I'm just saying the Prescott scales up more if you can control the heat and if you can overclock the snot out of it. At 3.8 to 4 gig, the Prescott starts beating up on the Northwood.

My comparisons of the 2.8C vs. 2.8E are perfectly valid since I used the same tests with the same OS with the same system with the same components except the CPU. However, I admit if you are using the new Sandra version, you can't make a comparison with your system vs. mine... but, you can't really do that anyway since the components are different.