Air cooled Celeron 2.8 to 4

Im almost new to overclocking, but was wondering if you guys could give me any input. Im running a celeron D 335 2.8ghz (prescott) air cooled on an abit ic7-g motherboard (geforce 6600 + ocz ram, overall my bargain bin mid performing system). So far ive brought it to 3.9 stable, but am worried I might do damage bringing it to 4. Idles at 43 and on PRIME95 load it runs as high as 50C. Too high temps? PRIME95 ran stable for more than a day. Any reccomended max VCore's for Celeron D's? To push or not to push...
(oh btw, ambient is 26 degrees)

first to the forums!!
and, 50C is not bad for load temps, some pepl have seen as high in the 60's for load temps, try it at 4ghz and see what that baby can do!

EDIT:
make sure that its running at 50C, not hotter. if you dont have an external probe, try the hot to the touch method after a few hours of load....

Temps are fine. You can run up to low 60's no problem. Vcore is fine up to about 1.45v or so. I don't recommend going above 1.5v on a socket 478 and only that if you have great case ventilation and maybe even having sinks on the mosfets.

batboy is wise...very wise. I wouldn't go above 1.5 under any circumstances, unless you went with phase cooling, and that's whole completely different ballgame.

4GHz is a 1.2GHz OC, and that's very possible. Intel's love to OC, and usually have pelnty of headroom unless you get the occasional dud, but those are few and far between, and it's already appearant that you don'y have a dud there.

I would try to keep temps below 50's though. I err on the side of caution with Northwoods becasue I killed a 2.4C chip, and it was obvious it was because of voltage (1.5v) coupled with too high temps.

If you want to err on the side of caution, I would look into water. With that, you can get load temps down and will have better stability even if it's not needed to achieve the 4GHz. Lower temps = better...***always***. In simple terms, electricity don't flow right when things are hot. The cooler you get electronic components, the less resistence there is, and the better the electricity flows, and it's also more efficient.

There are very good water cooling systems out there right now, and they really work well. Even if you are not looking to buy right away, gto into the water forum, and look around. There are some very interesting numbers in there, and it's really not that hard to set one up. They now have water blocks that are universal and can be used wqith many different CPU standards so they can be moved to another rig later if you need to.

The Prescotts can handle more heat than the Northwoods and the IC7 reads a little warmer than actual. So, 60 degrees load CPU temp is no problem.

Well, thanks for telling me the IC7's overread. Thats always good to hear! 1.5v Vcore max would be for the general board, so even if the CPU temp stayed very low over 1.5 I shouldnt even try? (Thermaltake Tower112 has proved its worth)

And to 3DFlyers response, I thought I had heard somewhere that Prescotts were at least slightly more voltage resistant than the infamous Northwoods and their NSDS. I guesss ill go check the water forum soon, but from past experiences the impression I recieved was that although water cooling could very well be the cooling of the future, it seems to me that overall with maintence, huge prices (pump + blocks = $$$), and horror stories it isn't worth it to me yet.

3DFlyer said:

I err on the side of caution with Northwoods becasue I killed a 2.4C chip, and it was obvious it was because of voltage (1.5v) coupled with too high temps.

Click to expand...

Must've been the heat, 'cuz 1.5V Vcore is pretty much dead on for the Northwood's stock Vcore range of 1.475 - 1.55V, IIRC.

Now start talkin' about getting much >1.65V w/a Northie, and yeah, I'd say GNDS has a decent chance of occuring over time.

But, yes, you are definitely right in that Northwood's do tend to be "somewhat" fragile, concerning overvolting, and taking care to not excessively overvolt a Northwood is important for it to retain it's ability to maintain it's OC.

'Course the OP is OCing a Prescott variation, so I suppose the Northwood specific specs are somewhat moot, but the idea of taking care to not overvolt, is a valid on.

Tho around here, it's generally been the Prescott that kills the mobo, & not the other way around!

Strat

Right, on a socket 478 mobo, which were not really designed for the heavy power loads of a Prescott, high voltage tends to kill the mobo, not the CPU. Look at my socket 478 Prescott cooling article link in my signature. There is a pic of my sinked up IC7.

Too bad... normal use its now at 38C and it doesnt break 50C in full load. I guess ill go add some sinks on the mobo for good measure. 1.5v is too low to bring out the real speed on my prescott. Im just going to have to head off to memory timings...