random bsods when OCing

when i OC my CPU, even a little, on my P5K-E and freezer 7 pro, when i game, i get random driver_irql bsods, linking to ATI drivers, but it only happens when i OC the CPU no GPU. but when i leave thibngs to auto, non clocked, it wont do it. im sure its not heating up, and i ahve cleaned the cooler.

i put volts at 1.3625 on my BIOS, and under load (on CPUZ) it fluctuates between 1.336 and 1.344.

my NB is set to 1.4 and RAM is 1:1 at 375 at 4 4 4 12 at 2.1

and my multi is 8.

my PSU is a second hand tagan 430W U15.

i have done memtest for over 5 hours with about 7 passes (twice)

so i am sure its not a ram problem. i have swapped gfx cards, and still get it, and switched to nvidia and get it with nvidia driver equivelent.

my pcie frequency is set to 100, but eevn if i set it to 110 it still happens.

, i have also tried it with 1.4Vs at Vcore, and 1.55 at NB, but still happens. i have loosened my timings to 555 15 and still nothing.

ram is set to 2.1

Your PSU might be struggling... Does it BSOD at stock when you prime?
Are there any power hungry stuff in your pc?

well apart from my x2900GT, nothing. runnign one HDD and one ODD. at stock nothgin fails, runs with out a hitch, but the secodn i OC, it craps out. i have ruled out RAM, and i know this chip can hit 3.45GHz (albeit under water) yet under air, whiel gaming, it doesnt want to work :(

You may well have just hit the limit.

Quote:

---

Originally Posted by Blitzen

You may well have just hit the limit.

---

I doubt stock is the limit... no system can be that bad (unless the mobo prevents overclocking)

You need to remember that al the components are rated at a given speed/voltage for a reason. Mfrs don't set out to make CPUs of different sppeds (within the same production line) but manufacturing tolerances dictate different performances - and the processors are graded accordingly.

So as soon as you start running them faster tan the tested specification, you are running into the risk of instability. That might be overcome by increasing the operating voltage - which in turn increases heat diissipation and internal stresses, and will tend to shorten component life.

It is true that many people overclock without significant problems, and a lot depends on the mobo set up - but to re-iterate - once you go outside those tested parameters, it is a bit uncertain. You may well have reached the limit by running your particular selection of components on that motherboard and the rated specification. Changing (for example) the memory might 'improve' thgings - then again it might not.

It might be worth running memtest+ both at stock and at overclocked sppeeds to see how that is performing though.