Overclocking (and volting?) my conroe

[The I]

Hi all, I just received my E6600 yesterday and after burn-inn I've started experimenting with oc'ing.

So far I've reached my first 'checkpoint' 3.06 ghz at 340.46x9 on stock voltage which was my immediate goal and a 25% oc I could be satisfied calling it a day with.... But of course I'll try to go further, just for the sport of it

Now the thing I'm uncertain about is whether upping the voltages is a good idea in my case. As far as I've understood (from a post at the xtremesystems forums) overvolting will have some long term damaging effects on my CPU (which at first will lower the max frequencies but at length could make it unstable).
This worries me because I'm going to be keeping this system for a long time. My last system was a northwood A 2.0 ghz with 512 mb pc2100-memory and a gf4 Ti4200 and I've had it since summer 02. It is quite probable that this system will last as long as my last one so I won't be risking wearing out my CPU.

On the other hand, I am not one to run with my PC on 24/7 either. A roughly estimated avarage would be somewhere around 3 hours a day over a year. Am I correct in guessing that in that case I'll wear my CPU down as much in 4 years as a guy running it 24/7 in 1 year?
And another factor I'm not sure about: I'm using EIST, does this lower voltages when the system is not stressed? - and in the case it does, am I correct in assuming that overvolting will only wear the CPU in stressed situations in that case, where EIST doesn't underclock and undervolt?
I would guess that my CPU isn't under stress more than 50 % of the time. I probably spend less than half my time playing games or doing statistical calculations, the rest of the time it will just be surfing the net, writing assignments in word and excel and playing music...

So:

1. Is it correct that overvolting will wear my CPU?

1. How much of a problem will that be?

1. Do you have a maximum recommendations of how much vcore I should do, in order not to damage my CPU during it's expected time of use?

I might be hitting another barrier to though: at my current oc I'm hitting temps around 63 degrees C using StressPrime 2004 ORTHOS and Core Temp 0.9.0.91. I guess I'm hitting another barrier here, though I'll probably be looking at how to improve airflow a bit in my case (Lian-Li S80)

 

[gobucks]

this kind of stuff is open to argument. most people who overvolt their CPUs are the same kind of people who upgrade their PC pretty frequently. as such, there isn't a whole lot of definitive data on the subject. that being said, i think the general concensus is that minor overvolting won't significantly lower a chip's life, but major overvolting can. I think a general upper limit for overvolting is about 10%, which would mean about 1.45V for a conroe.

However, at 1.45V, my e6400 @3.24GHz using a zalman 7700, was running at a dangerously hot 68C. Intel says anything over 61C can lead to instability. Unless you have really good cooling, i wouldn't go past 1.4V. Currently I have my setup at 3.08GHz, 1.375V, and it hits around 55C load.

With your system, you're pretty much at the maximum without getting better cooling. 63C is pretty high, and you haven't even bumped up the voltage. what kind of cooling are you running? Until you get something better, you may want to try opening the case door, or lowering your clockspeed slightly and seeing if you can undervolt your CPU (not much, but even .01V could make a difference)

 

[ForgetCassettes]

gobucks, what program are you using to determine cpu temperature?

 

[Some1ne]

Is it correct that overvolting will wear my CPU?

Yes.

How much of a problem will that be?

How often do you upgrade? My X2 3800+ has been running at a pretty significant overvolt (right at that 10% threshold, I think) since the day I got it (I think roughly a year ago, or something like that), and prior to that, my plain Athlon64 3000+ was running at a slightly smaller overvolt for the duration of its life, and I've never had any problems with either of them. Basically, unless you plan on keeping your CPU for more than a couple of years before the next upgrade, there's probably not a whole lot to worry about.

Do you have a maximum recommendations of how much vcore I should do, in order not to damage my CPU during it's expected time of use?

Depends on how longyou expect to keep the chip around, though anything within 10% is generally considered to be okay. Above that maybe if you have very good cooling, or if you upgrade your chip frequently.

 

[The I]

Originally posted by: gobucks
this kind of stuff is open to argument. most people who overvolt their CPUs are the same kind of people who upgrade their PC pretty frequently. as such, there isn't a whole lot of definitive data on the subject. that being said, i think the general concensus is that minor overvolting won't significantly lower a chip's life, but major overvolting can. I think a general upper limit for overvolting is about 10%, which would mean about 1.45V for a conroe.

However, at 1.45V, my e6400 @3.24GHz using a zalman 7700, was running at a dangerously hot 68C. Intel says anything over 61C can lead to instability. Unless you have really good cooling, i wouldn't go past 1.4V. Currently I have my setup at 3.08GHz, 1.375V, and it hits around 55C load.

With your system, you're pretty much at the maximum without getting better cooling. 63C is pretty high, and you haven't even bumped up the voltage. what kind of cooling are you running? Until you get something better, you may want to try opening the case door, or lowering your clockspeed slightly and seeing if you can undervolt your CPU (not much, but even .01V could make a difference)

You're using core temp to measure your temps, right

I see two or three possible explanations of my fairly high temps (just checked: on normal frequency it does 59 degrees C under SP2004).

As far as I've seen in reviews my Scythe mine cooler should perform pretty decently.

I see three possible causes:
1. Might it be that my thermal paste (artic silver 5) haven't quite settled yet? - I've burned in for about 12 hours and probably run the computer for around 10 hours apart from that.

2. The case I use (lian-li S80) has a pretty odd airflow path, it pulls the intake air in from vertical openings in the back of the side panels, then the air is drawn on the inner side of the outer side panels (it has two separate sets of side panels) to the two front mounted 120 mm fans and into the system. Apart from the two 120 mm intakes it has a 120 exhaust 'behind' the cpu (by standard it's set as intake, but I figured it would work best this way) and the PSU fan.
I realize this is far from optimal - I bought the case for the looks and the build quality, not for optimal cooling - and it probably has something to do with the result. The strange thing however is that removing the side-panels actually makes the CPU hotter!

3. I might be able to get the fans running at a higher speed - haven't quite figured out how the Asus fan-controller works.

I tried lowering the voltages (the smallest step I could take was to 1,27 v) and it still seems stable at my current oc - have been running SP2004 for 10 minutes now. My temp is at 60 degrees now.
I guess this meens that temperature is the limiting factor for me since I could probably oc a bit more still on stock, since 1,27 seems to be enough at 3 ghz...

 

[OBCT]

I believe burn-in time for AS5 is 200 hours. Just leave your computer on, run Prime95 constantly, whatever, for a few days until the AS5 gets kicking.

 

[gobucks]

Originally posted by: ForgetCassettes
gobucks, what program are you using to determine cpu temperature?

i've been using Core Temp to monitor my temps. it's possible that it's misreading the temps, but i really don't feel comfortable assuming that. in gigabyte's Easytune software, it read like 58C when Core Temp said it was 68C. Oh well, 3.08GHz is enough for me right now, maybe if i get ambitious some time down the road I'll get a scythe ninja and try to squeeze more out of it, but not right now.

 

[The I]

I've been experimenting some with undervolting. I haven't tried any lower than one step below max at my current oc, but I tried stepping down the fsb to standard value and see how far I could get. 1.225 vcore seems to be the minimum value on my MB, which is what I seem to be able to run stably now (so far I've primed 12 minutes sans errors).

I also tried experimenting with speedfan to find an idle / surf the web setup. Right now I'm running all fans on 45% which is very hard to notice over my ambient noise level, it's priming at 60 degrees on core at my current undervolt. Since this setting is for internet and playing music (maybe dvds to) I presume that's ok, sine full load isn't a realistic scenario.
This setting idles at around 51 degrees so far (might go lower if I'm patient enough to wait around longer), but in speed-fan I can see that the reading on the other temperature probe (the one not in the cpu) is actually substantially higher (46 degrees while the cpu is at 36), so I guess I have some warm-air buildups around the case, should I consider that a problem?

I hope that my AS settling will help me win a few more degrees and then I'll start working on a higher max-overclock (probably on stock voltages, so far I seem to be able to go pretty far on an undervolted setup).
I'll probably also be swapping the 100 mm fan in the Mine-cooler for a 120 mm one (I've seen tests showing that'll be more efficient), I'll be looking for something that has a higher max CFM at the cost of higher max db since the 100 I have now seem to be able to go lower than I need it to already. I hope that will shave of a few degrees to.

Thanks for your advice so far. In spite of the small heat-problem I simply love this system!