help me overclock my p4

[spy2520]

well i currently have two systems, one is a pentium d 930(presler), zalman 7700cu, gigabyte g1975x, pqi ddr2 pc5300, x1900xtx, all the good shit.

the other is a pentium 4 HT 3.0E(prescott), thermaltake tower112, asus p4c800-e deluxe, dont remember the ram ( ithink its kingston value pc2700) 9800 pro, all the older almost good shit.

anyway, when i overclock both machines i run into a wall around 242 MHz fsb. for the asus i can see why this is taxing, but the gigabyte should support the p4 EEs and a fsb of 266. i tried increasing the vcore on both but i dont know what the limits are so i only go a little over. i also messed with the ram multipliers to keep the memory stable, the pc5300 is only operating at 646MHz and stock timings. the pc2700 is operating at around 320 MHz. i also tried changing the vdimm and fsb voltage on the gigabyte (just a little) and it didnt make much difference. of course the prescott is several degrees higher than the presler at idle and on full load, but neither seems so hot as to cause instability.

so far i got the presler/gigabyte up to just under 1GHz fsb(qdr) but the system would run for 30 minutes or so and then reboot, all the programs that measure cpu temp are reading either really low temps (i.e. -55*C) or not registering a temp at all, so i cant really measure the temps for the presler although i'd guess it would be somewhere between 40-46*C at idle.

 

[spy2520]

bump ttt

 

[Eastcoasthandle]

you will need to put a fan on your passive HS for your chipset. Also place mc14 on your mosfets. Then download and use Memset for your mem. MemSet 1.X


Usually a wall like that is the result of something overheating...usually those mosfets or your NB chipset which I assume you already have OC.

You will also need to bump up the voltage for your VCore, incrementally at a time. I have a 660 OC 4.36 @ 1.5000V before it was OC to 1.5375V and I knew something was wrong because that was way to much voltage for the Vcore. Once I but screweed a fan on my NB chipset I was able to turn down the Vcore to 1.5000V and it's Prime95 Stable. So that should tell you something right there. I am not sure about your other chip because I don't use it. But I assume the same will apply but with less voltage.

One thing I do know about the Preslers is that they require the 8 pin connector in order to properly OC. If you have a 4 pin connector from your PSU on the 8 pin connector on the MB IT WILL NOT OC PROPERLY.

Next once your realize those results are nominal its time to go water buddy...it's the only way to go.

 

[spy2520]

the gigabyte motherboard has fans on the mosfets and northbridge already. it has four of them, i'll try and find s pic online.

 

[QualiC]

I'm not clear on when the temp readings are screwwy, at moment of crash or prior to that? Would've assumed its just at point of freeze-up, except if this is a predictable/recurring event, I'm sure you've sat and watched the reading closely from like 20min till crash, and would thus know if heat spikes on any component?

Could main components be staying cool enough to be stable, but case temp getting hot enough that its causing something like a HD to fail?

Q

 

[spy2520]

the motherboard never really reads the cpu temp correctly fomr what i have seen, or rather it reads correctly in bios but in other programs such as sandra and systool it comes up as -55*C all the time.
the case temp hovers around 32*C and never really goes over that, i also have one of the sides panels not on the case, just kind of propped up next to the case so the fan in it blows towards the video card.

damn they're small, and i dont know why, i saved them full size.