Recommened voltages for 1st overclock

[DJ898]

Started my first oc. My voltages for right now are set at auto however i have read that as you increase the fsb the voltages increase as well which makes me a sad panda So I was wondering what voltages should I use? As starting points because right now CPU-Z is saying that I am using 1.392 V which is about the limit right for a Q6600 G0? My specs: Q6600 G0, Maximus Formula, 750 Watt silencer, 4gb xms2 800 MHz ram, ultra 120 extreme. Now these are voltages for everything the nb, sb, ram etc. I just don't want to blow up my board
Any help would be appreciated but you guys have been an awesome help with my first build

 

[xringx]

I would manually set them all to stock voltages, and then work you OC. If you dont know the stock voltages, look them up.

An reason why its important to know your stock voltages is for example on my MB, stock dimm voltage is 1.8, but my bios doesnt say that, it only shows +'s or -'s to that. I had to look up the standard voltage. My memory, runs at 2.2v standard.

Depending on how much OC you get, you can start tweaking your voltage just a bit to see if you get the system to boot/run stable.

Its a long process and every computer is different and each OC require different tweaks at various steps.

For example, I can run my system with super low vcore of 1.125 at 3.0Ghz. To get to 3.25 or higher, I really had to move to near stock vcore. Its really a big move in terms of vcore for just to go from 3.0GHz to 3.25Ghz

 

[DJ898]

cool thanks. However I just ran Prime95 overnight and all my cores were stable until core 0 got a fatal arror after 11 hours 36 minutes. Should I be worried or should I just leave it?

 

[xringx]

I wouldnt worry about it. 11hrs is good. I never manage to get over 8hrs

 

[Groveling_Wyrm]

a Q6600 is safe up to 1.5V with good cooling. Running anything more could be dangerous.

DDR2 voltages range from defaults of 1.8V to 2.2V You shouldn't overvolt your memory by more than + .1 volts unless you are sure it can handle it.

Your Asus board is actually a pretty versatile board, and can take a lot of voltage tweaking. Just make sure you monitor your temps while doing any voltage changes. A severe rise in heat is a good indicator that you overdid something...

 

[DJ898]

Yeah I just finished the oc on my cpu. Up to 3.6 Ghz Luckily enough I 'only' have to use 1.5 volts But I can't go any higher and a 50% oc is good enough :lol:

 

[Groveling_Wyrm]

If it won't go any higher, I would look at your motherboard voltages or possibly ram settings. That is usually what holds the OC back at that speed.

Another tweak is that there is sometimes a "Dead" zone where it will not work within a certain range of FSB settings, but it will work fine above and below those settings. For example, your motherboard will work perfectly at 340 FSB, but not at 360 or 380, but will work again at 400.

You might want to CAREFULLY check and see if that is the case. If you do this though, read up on your voltages first.

 

[systemlord]

cool thanks. However I just ran Prime95 overnight and all my cores were stable until core 0 got a fatal arror after 11 hours 36 minutes. Should I be worried or should I just leave it?

If you get a fatal error anytime thats not a good thing! If you can't run Prime95 for 24 hours without errors its not stable at all. "xringx is mistaken about the fatal error being ok." If your motherboard is an X38 then heres what your stock setting should be. If your at 400 FSB like I am the Northbridge is at 1.41v and my CPU Vcore 1.33v with all other setting at stock. Heres a review of the X38 chipset.

Xbit Labs http://www.xbitlabs.com/articles/mainboards/display/asus-p5e.html

By systemlord at 2007-12-17

 

[systemlord]

I wouldnt worry about it. 11hrs is good. I never manage to get over 8hrs

NO, NO, I would worry about your Prime95 showing a fatal error at anytime! This right here is bad information, sorry xringx but this is not right. You think that a fatal error is ok? I am shocked by this!

 

[cnumartyr]

NO, NO, I would worry about your Prime95 showing a fatal error at anytime! This right here is bad information, sorry xringx but this is not right. You think that a fatal error is ok? I am shocked by this!

I think 24 hours is excessive.

8 Hours Small FFT, 12 Hours Blend with no errors and it's fine.

I've never had an issue with any system that passed that test even having been left on for weeks at a time.

 

[systemlord]

I think 24 hours is excessive.

8 Hours Small FFT, 12 Hours Blend with no errors and it's fine.

I've never had an issue with any system that passed that test even having been left on for weeks at a time.

Again I am shock by how you can say after 12 hours is enough! Even when the software program says that if your computer can run with no errors for 24 hour, yet you make up your own rules completely disreguarding the instuctions givien to you by the software designers.

8 Hours Small FFT, 12 Hours Blend with no errors and its fine.

You think 24 hours is excessive, but the software designers think 24 hours is not excessive. We need to stop thinking and start following directions, instead of making up our own. The OP did get 1 fatal error, and thats not ok if you follow the instuctions.

 

[cnumartyr]

Again I am shock by how you can say after 12 hours is enough! Even when the software program says that if your computer can run with no errors for 24 hour, yet you make up your own rules completely disreguarding the instuctions givien to you by the software designers.

8 Hours Small FFT, 12 Hours Blend with no errors and its fine.

You think 24 hours is excessive, but the software designers think 24 hours is not excessive. We need to stop thinking and start following directions, instead of making up our own. The OP did get 1 fatal error, and thats not ok if you follow the instuctions.

I didn't make up my own rules, I spoke from personal experience. In my personal experience I've found no difference in testing for 24 hours of Small FFTs versus 8 hours of Small FFT + 12 Hours of Blend for stability of daily usage.

I didn't say 1 Fatal Error was ok. I said in my opinion 24 hours is excessive. I really don't care what the software manufacturer says. I used their guidelines for a while, and now I've found my own that work. I haven't had any issues with this system that failed a 2.5 hour Prime95 Small FFT. I bumped up the vCore 1 notch and it ran 8 Hours Small FFT followed by 12 Hours Blend without an error. It has now been on for 2 weeks without a shutdown or restart running F@H in the background as well as encoding and playing games.

I apologize if my standard isn't as good as yours, however it works for me and hasn't led me wrong yet.

 

[systemlord]

I apologize if my standard isn't as good as yours, however it works for me and hasn't led me wrong yet.

There not my standards there the standards provided be the software designers, why do people think that there my standards? Can people not read correctly anymore?

 

[cnumartyr]

There not my standards there the standards provided be the software designers, why do people think that there my standards? Can people not read correctly anymore?

My bad, semantics. I'm sorry my standards aren't as high as the developers of Prime95.

When my computer crashes and is totally unstable I won't be here crying about how it passed on my standard. But hey it hasn't.