Legit Reviews

First off, this processor has been overclocked to varying speeds throughout its life. It was run at 2600MHz for almost a year before the K8N Diamond Plus underneath it died (the PCI lock failed and the overclock killed the PCI slot with my X-Fi, X-Fi is OK but I couldn't get sound out of the mobo, not even the onboard). It was over-volted at that point and had a Scythe Ninja Plus on it.

The 2.6GHz overclock was not "Prime Stable" but it did absolutely everything I threw at it without ever showing any signs of instability (until the mobo failed). I replaced the MSI with an Asus A8N32-SLI Deluxe. I tried OC-ing again, but no dice. Even 2.2GHz was causing Windows to freeze up after a random period of time.

I undervolted my CPU just for fun. Two days later I bumped the volts back up to one notch under stock instead of two notches below (1.35v, my mobo was giving it 1.4v by default for some reason...this Opteron was designed for 1.35) and tried 2.2GHz. Twelve hours later, my PC has not frozen, hung or crashed.

I've played the Crysis demo for an hour, run SMP F@H constantly after that and transcoded some AVIs using Nero Vision for a good six hours (two DVDs worth, it is not multithreaded so it is slow). No signs of instability. No freezes. No crashes.

What is going on?

The CPU fairies paid you a visit.

Honestly, who knows. Maybe the undervolting caused it to be conditioned to lower voltage. My gut feeling is maybe the quality of power coming in could have something to do with it.

Dan

Maybe your replacement motherboard is just a dud. I've had three different 939 boards and they all had different quirks and overclock limits. It's all luck of the draw really.

My replacement is working better than the original, seeing as that I don't need to overvolt at all to go higher than stock, actually I'm below stock volts and above stock speed. That's why I'm confused.

most people dont understand ohms law, when u increase the voltage, you also increase current, when u have a cpu runing @ higher voltage it will result of circuit trace burn in, what it mean is, it makes the pach where the eletron move wider due to the extra voltage/current, En = -1/n^2 (me^4)/(8E^2h^2), every time u increase a uV it increase the path by 2.492E-32 m, when the path is too wide and too close to the other path, the eletron will jump over to the nearest or the lease resistance path, as result short cuircuit.

example is intel's prescott, the heat is from current/voltage leakage, the fix is wider path and alot fo grounding.

ur case, RMA the CPU

The CPU isn't broken, it was performing beyond and better than spec before. Now, I rebooted and WHAM the system destabilizes. So I'm back to undervolt-stock...

Hopefully the computer gods will let me run 2.4 next week