Guest
CompuDav
2004-05-06T19:03:58Z
Once upon a time, there was an Intel Pentium 4 CPU in an ASUS P4P800 motherboard. It was running nicely at its rated 3.0 GHz until one day its owner decided to overclock it. The owner had read many stories about people reaching 3.6 GHz with their 3 gig P4s, so he decided that it would be worth overclocking. He slowly raised the FSB, checking for stability and found that he could hit 3300 MHz without any VCore increases! So little by little the owner pushed the FSB until it was at 3345 MHz, where it wasnt stable.

OK, the owner thought, Ill just increase the VCore by 0.0125. The CPU was stable until it reached 3360 MHz, where it wasnt stable again.

OK, the owner thought again, Lets try another 0.0125 VCore increase. Magically, the CPU was stable. Then the owner bumped the FSB to 225, or 3375 MHz and it wasnt stable again. Another VCore increase was given, making it 1.5875V. The CPU was fairly stable, however, no matter what, no VCore up to 1.6V would make anything over 3375 MHz stable. The owner even went so far as to try an Extreme Spirit North and South Bridge Cooler and an Antec 550W PSU to see if that would help.

For those of you who are confused by or too lazy to read the tale, heres the bottom line: I cant overclock this no-good, uncooperative, P.O.S. CPU over 3375 MHz (225 FSB) without stability problems even at 1.6 VCore. My goal is/was to get it up to 3.4 GHz (227 FSB).

All temps are fine - CPU is always under 50 C, and the mobo is always under 35 C (usually 30-32 C).

My questions are as follows:

a) Is it usually necessary to put the VCore above 1.6V for a 400 MHz increase on a 3.0 GHz 800 MHz FSB P4?

b) I tried it at 1.625V but noticed that P95 failed in 1 minute. Granted, I gave it 0 burn-in time, but even still, one wouldnt think that it would fail so quickly after being bumped up from 1.6 to 1.625V. Should I try this with some burn-in time?

c) Am I just stuck with a no-good, uncooperative, P.O.S. CPU and mobo comination. It is a week 9 CPU which I thought were pretty good for overclocking. I think week 10 was amazing and the weeks surrounding it were pretty good as well.

Any suggestions?

Thanks in advance,
David

P.S. Yes the fairy tale format is kind of weird, but you don't know how much mental anxiety this has caused me. It makes NO SENSE!!!!!

P.S.S. Sorry if I posted in the wrong forum - I assumed since it was relating to a specific processor that it should be in the processor thread.

:banghead:
Jordan
  •  Jordan
  • 50.6% (Neutral)
  • Advanced Member
2004-05-06T19:23:08Z
There's one thing you've neglected. Overclocking is an unknown quantity! A lot of it is based on luck! 3375MHz is a decent OC and it doesn't mean your PC is crap, you were lucky to get this far!

What is your RAM and CPU cooler. If it's 3200 RAM it's very much likely to be your limiter if it's not the CPU. Try running with the RAM with a lower devider and OCing to see if that's your limiter.
CompuDav
2004-05-06T19:29:03Z
It is 3200 RAM but it is running at a 5:4 ratio so I think it is OK. The RAM has no cooling but the CPU cooler is a TT Spark 7.
Jordan
  •  Jordan
  • 50.6% (Neutral)
  • Advanced Member
2004-05-06T19:37:36Z
Yeah RAM should be OK at that, though I only recommend 1:1 ratios for OCing. You're limiting your FSB to 180 (720)MHz that way. I never understood high clock speeds with low while limiting the FSB. It just creates an FSB bottlenack.

RAM cooling really doesn't make a difference though if it's in the line of good airflow I beleive it OCs better. I believe this because my HSF is so big it overlapps the RAM and northbridge and I get better FSB OCs now than before I had it.