3.0 Prescott - Voltage?

Got my new 3.0E. Its running Prime95 testing the L2 cache right now. Its at 55C on the stock heatsink, and no case fans. It seems to have stabilized there now. Pretty nice I think.

http://processorfinder.intel.com/sc...cFam=483&PkgType=ALL&SysBusSpd=ALL&CorSpd=ALL

And of course I'll list how well it can OC and such later.

What voltage can I go upto on this? I won't go above 65C load though. I'm at the lowest my mobo supports. Its fluctuating from 1.25-1.35v right now.

Seems most people say to stay below 1.5v on the prescott's. I'm running my 3.0E at 1.485 right now, 245FSB.

I run mine at 2 differnet volts.

3.75ghz, 250fsb, 5:4 divider, Ill use 1.55, which is actully around 1.5.
3.45ghz, 230fsb, 1:1 divider, Ill use stock voltage.

Both are stable for days, full load and everything.

I didn't go above 1.4. It was too hot as it is. Its already for sale. lol

check this out for max voltage

http://users.erols.com/chare/elec.htm

You say its too hot and you have no case fans!! i wonder why?
My 450mhz k6 gets very hot without a case fan, so a 3000mhz pressie is gonna get a tad warm.
Mine idles around 33c and around 50c full load laterly

My 2.4C does 33C-49C without case fans at 250FSB on a stock heatsink.

But I also tried it with case fans, didn't do too much. It just emmits too much heat in general, and didn't seem as fast in windows in general.

yeah I ran mine close to 1.48v and killed my power supply.............500watt. just smoked it.so running at higher vcore does do other damage........take heed.

Xp got a huge speed boost when i did the cache reg edit setting from 0k to 1024k.
As windows xp is not able to set this automaticaly for some reason.

When talking Northwood, we usually say that "Don't go over 1.7v", mainly due to the SNDS cases. Tomorrow I'll get my 3.2 scottie for some benching with highend watercooling. The cooling performs good if I may say so myself, it can handle 4GHz Northwood primestable, and I hope to cool down this chip at atleast 3.9.

Is 1.5v the limit for 24/7 use?

Thanks for all answers.

Fredrik said:

Is 1.5v the limit for 24/7 use?

Click to expand...

The CPU's seem to do fine but it is the motherboard's MOSFET's and your PSU that gets strained from this. I beleive you do not want to go over 1.5volts for 24/7 use. I do not have first hand Prescott experience, so I would wait for some other people to tell you before you start bumping up the vcore

Fredrik said:

When talking Northwood, we usually say that "Don't go over 1.7v", mainly due to the SNDS cases. Tomorrow I'll get my 3.2 scottie for some benching with highend watercooling. The cooling performs good if I may say so myself, it can handle 4GHz Northwood primestable, and I hope to cool down this chip at atleast 3.9.

Is 1.5v the limit for 24/7 use?

Thanks for all answers.

Click to expand...

Yes for Prescotts people are saying not to go over 1.5vcore,my 2 pressys I keep under 1.5v........... You mosfets and p/s really take a beating.........my mosfets are all heat sinked on my 2 C7's (running prescotts on them)
I gave my 3.2E too much juice last week and smoked my 500 w. p/s........ just replaced it and off running again.

Dual Prime @ 3856MHz and 1.475vcore
Pics coming up later tonight.

My personal experience is a little different than most here. I got one of the first 2.8Es, and she's a dog. Only does about 3.2GHz on default vcore. But it responds well to increased voltage. I run mine 24/7 at 100% load (2xF@H) at 1.525v for 3.4GHz. Haven't had an issue yet. The P4C800-E hasn't shown any stress from this, mosfet temps aren't ridiculous and sinking them didn't improve temps (I use thermal probes to check mosfet surface temps). CPU temps stay around 43-45°C at load (SP-94 + Delta HHE), but this is very dependent on ambient temps (I keep my place as cold as I can, around 65-68°F). I'm using a PC Power & Cooling 510 TD, and the 12v rail doesn't even show much of a hit at these voltages. It didn't show any reduction until vcore went over 1.5v. It's probably good advice to stay under 1.5v, but it's not a hard fact that higher voltages will kill some component in your rig. Cooling it is a more important issue, IMHO. Good luck.

IMHO in having observed things from an AMD perspective, you FIRST want to get it below 40C THEN you can try more voltage, that's even if you need more voltage when you keep it cooler. Intellers seem to be used to really robust chips that you could let get up to 70C without a care. I don't think this is the case any more. Pretend you've got an AMD and baby it. If tbred experience is much to go by, once you've topped 45-50C, you have to throw in large chunks of voltage for very little return apart from dead chips. I mean, 200Mhz sounds like a LOT of gain, and it was, back when we were mostly below 2G. it was 10-15%, now, people are like killing their chips for like 6% more than you can get on stock voltage. At current chip prices, that extra 6% seems to cost you guys about $100 a month.

If you wanted the speed for Distributed Computing for example, you could take that $100 and get a barebones crunching/folding rig that would add 25-50% to your output, and do the same the month after..... it's an expensive 6% for DC.

And by the time you've killed 2 CPUs, you could have had a duallie rig... if you can all your OS processes running on one CPU and your game on the other, you're probably getting 10-15% better than that CPU in single CPU config.....

I know the numbers look big and shiny, and you think they could be yours, but stop one moment and think how much that all too brief rush will cost you. We didn't used to care as overclockers, when our CPUs "only" lasted a year or two, there was something MUCH better by then, but now the burnination cycle has come down to 3 months, less maybe, and the release of new products looks like slowing, and the prices look like holding, this attitude can't be maintained along with your sanity or bank balance.

Back in the good old days we were getting about 10% more overclockage over stock voltage overclock for 10% gain in voltage, then your next 10% would get you maybe another 5% with reduced life and your next 10% would get you another 2.5% if you didn't get a dead chip real soon. I think these CPUs are coming at a level one overvolt as it is. Another 10% on top is getting into the danger zone.

Damn, I ranted


Road Warrior