Using folding at home to test stability.

[Peter Trend]

Just a quick question to any of you expert overclockers who know the odd bit about folding@home...

Can I use F@H instead of Prime95 to stress my computer after an overclock, to determine whether it is stable or not? Are there any areas that will not be stressed enough to declare the OC as stable?

Thanks

 

[Toadster]

I use BOINC at work to burn in new servers for at least 24-72 hours... seems that if the boxes are stable running 24x7 on that - then they're pretty good for everyday work.

 

[Peter Trend]

Thats what I was hoping to hear, so I finally got 2.38GHz stable on stock cooling on the Phenom 9500. That's 80MHz more than I got on the Gigabyte MA790FX-DQ6 board before I fried that. But the 9500 is hard to overclock, it dies on high FSB and I cant up the multiplier

 

[aigomorla]

do you really want to?

You want to fold when your overclock could not be stable and find out you failed a WU later on?

You should ask Markfw. If anything he could probably tell you if you should or not.

 

[Peter Trend]

Well...basically it doesnt really matter that much because my Phenom II 940 arrives on monday

But I did have a reason, because I dont want to waste 12 hours of my cpu time running prime95 when I could be folding. Luckily I havent failed a WU yet because of my overclock, only by messing with F@H files (moving them onto SSD in some vague hope of an improvement lol)

I expect the PPD from the Phenom II to be fairly disappointing without a big OC, and for that one I will try more stress testing

 

[Markfw]

Here is what I do.... Overclock and keep raising untill it won't even load windows stable, back it down a little. Then use superpi, and I always have to take it down a little more. Then go down a couple mhz from that, and try folding. If I get an EUE, turn it down, or crank the volts up until it appears stable.

Its my ultimate stability tester. If I can make it 30 days with no EUE's, then I am good. I may not have the highest OC's here, but I use the cycles to good use while I am testing. Usually me fist try is within 100 mhz of the final number.

 

[Peter Trend]

Thanks for that, I just failed Super Pi 32M, will back it off a couple MHz and try again.

 

[Markfw]

Superpi is fast (I use 1m or 2m) but doesn't catch problems. It just gets you close....

Then I use F@H.

 

[geokilla]

I'd just OCCT for 2 hours, then fold for as long as I can. Of course, there's people like Mark here so he's basically stress testing 24/7, with like those 2 min gaps between WU so the CPU can "rest" a bit. lol.

@Peter. When your Phenom II arrives, wanna post some comparison stuff compaerd to your Phenom 9500, as well as post some pics of the new stock cooler? I'm interested to see if it does indeed come with 2 heaptipes or not. Also, maybe it's time to switch your stock cooler out for a better one like the Vendetta 2 so you can clock it even higher.

 

[Peter Trend]

Yeah I can post some pics when it arrives. I'm already looking at water cooling the Phenom II, I do own a 120mm fanless heatsink from the last motherboard, but I had to bolt it on as I didnt have an AM2 adaptor. I think I will just leave it low until I can WC.

I'll make a few benchmarks of the 9500 as it is now at 2386MHz and save them for later.