Hi,
I'm not sue this where to ask... apologies if not, but CPU folds most of the day...

So I increased the FSB on my e7200.... now CPU-Z says it's clocked at 3.2GHz ( )
Cool.
My mobo is on Automatic for voltages and all the rest but FSB... I saw values changing as I was increasing the FSB Frequency.
I folded with the processor clocked like this all day yesterday, no problems.

My question is, do I need to change anything else in the BIOS to make it "more" stable ?

[EDIT]
CPU-Z screenshot:

OK, I'm replying to my own thread...

After looking around the internet, it seems that my VCore is a little high... But since I'd like to avoid toying around too much with the fine settings of all the voltages and frequencies, I'll leave it at that.
Also, going from 2.5GHz to 3.2GHz (30% increase), is good enough for now. I'll stick with that.

well the ultimate test is stability and temperature.

have you tested for stability with prime95 and occt? are your temperatures out of your comfort zone? that's what really matters here. whether your numbers perfectly match up with other random people on the internet shouldn't be that high on your list

If your motherboard operates like most what happens is the northbridge has a divider circuit and changes it's own multiplier at every even multiple of 66MHz. If your chip is designed for 266 MHz it will operate best at 333 or 400MHz. It may not operate at all from say 450 to 465 MHz, and then start working again at 466MHz.

You can also lower the multiplier and adjust FSB that you feel is best. My E6400 (266 * 8 ) total 2.13GHz, I run it at 7 * 400 for 2.8. It will run at 8 * 400 fine with extra voltage, which I would rather not do. I don't even have any heat issues, it's 65nm.

I would think you should be able to run 8 * 400 at stock voltage.

For stability and accuracy you need to run Prime 95 on each processor, so for you that would be 2, and I think there's a setting for which processor.

If you left that C1E option on (or whatever it's called in the BIOS) the processor may lower it's multiplier to slow down and use less power when it's idle. The makes for very strange results in CPUZ. All you need to do is to move the mouse real fast and the multiplier will change right before your eyes.

Really? I thought the ultimate test was beating Cordis.

I wish....

And yesterday I got 2 of those 47 points WU ( ) and one of those 88 points one done on the CPU... ain't it great ?

Thank you for your inputs.
I'll run Prime 95 this week-end as the ultimate test.
My "super duct" project was a miserable failure so I'm back to my previous airflow setup. It'll be a good test for that too. I'll post temps when I'm done.

Well, hope it works out, I don't OC much as a rule, I have one gpu that's overclocked, but it's only because I put a big aftermarket cooler on it. So good luck, until we get point updates again things are going to be kind of dull anyway.

So Frenchie, when will your E7200 start running 333 times 10?

Hi, I just read a very nice OCing guide on anandtech. Very interesting and detailed. So now I'm thinking of reducing the multiplier to 8 and increasing the RAM speed to 400MHz (full blast for the RAM).
If that works, I'll lower the voltages a bit if possible. I noticed the RAM sticks were hot the other night so I'll have to check those as well.
We'll see.

@aristide : I'll have to look, but I'm not sure I can raise the multiplier to 10. I think it's better if I increase the RAM speed first.

Frenchie,

Did I not suggest to you to run at 400 * 8 a while ago?

You do have to be careful though. If you board supports 400MHz native (runs at 400MHz for the processors without OCing) then you will be fine. But if it OCs to 400 only because you told it to, then certain tools, like Windows Update, may not work correctly.

I found that updates are successful when teh system is returned to it's non-OC state, install Windows updates, and whatever other software, and then return to OC. I also run disk defraggmenter without OC, I don't want any extra trouble.

Aris

Hey Aristide,
You sure did, except now I understand what you were talking about and before I didn't
So I just did what I wanted (and what Aristide suggested from the beginning), and now I'm going to run Prime 95 and see how it goes.


[EDIT] : ran prime for 30 minutes, no errors.
Here is what CPU-Z shows now : (darn, I have the wrong pic... I'll post the screenshot tonight; Voltages are back to almost stock, multiplier is at 8 and bus speed is at 400MHz.)

[EDIT 2] : screenshot :

frenchie,

try running P95 for 24 hours or more to make sure. If it can make it that long then it's more than likely gonna be okay. The longest time with an error using that approach has been 23 hours and 16 minutes. Of course, I was running an e2160 at double its original clockspeed, but hey, it happens.