Overclocking problems with E6400, help needed

[Ryzalis]

I'm currently running an E6400 and P5B Deluxe, and so far I'm extremely impressed with the overclocking potential of my E6400. I have it undervolted to 1.1v, and even then it's still going strong at 2.712 Ghz (339x8). I didn't expect it to overclock that much at such a low voltage. I have a feeling I could push it even more without raising the vcore.

However, even though I'm able to run dual prime with no problems for 10 hours (or more, I didn't have the patience to run it for longer) at that speed, I get a "not exact in round" error in SuperPi 32M within a few seconds. I'm running my Corsair memory rated at DDR2 800 5-5-5-12 at DDR2 678 5-6-6-18, so I'm fairly sure sure my memory's not the problem. All voltages other than vcore (1.1v) and vdimm (1.9v) are on auto.

So what's my problem? I know it isn't the CPU or the memory. Does small FFT prime test motherboard FSB for stability?

[Ryzalis]

Still looking for an answer. I tried raising vcore to stock, but I'm still getting rounding errors in SuperPi. It's definitely not the CPU.

[AndrewZorn]

perhaps try... overvolting the CPU? i mean, it IS an overclock... see if like 1.35v makes no rounding errors

[GoriLLakoS]

You probably need more Vcore or more Vdimm..:D

Give a try to raise the chipset voltage for a little..:D

Don't try only prime to test the stability of the system...try snm too..:D

[Ryzalis]

perhaps try... overvolting the CPU? i mean, it IS an overclock... see if like 1.35v makes no rounding errors

I tried that already (see second post). I still get rounding errors.

You probably need more Vcore or more Vdimm..:D

Give a try to raise the chipset voltage for a little..:D

Don't try only prime to test the stability of the system...try snm too..:D

My ram is rated at DDR2 800 5-5-5-12 @ 1.9v, and I'm running it at DDR2 678 5-6-6-18 @ 1.9v. I shouldn't need more vdimm if it's running lower than stock, should I?

It may be the chipset voltage, though. Can anyone here with a P5B Deluxe confirm what FSB you overclocked to before you needed to raise chipset voltage?

[RangerXLT8]

It may be the chipset voltage, though. Can anyone here with a P5B Deluxe confirm what FSB you overclocked to before you needed to raise chipset voltage?

No two situations are gonna be the same. If you have not raised chipset(MCH) voltage yet, there's a darn good chance it needs to be bumped up. Go up .1V and that should give you some headroom.

Generally a CPU is not going to run a stable overclock when it's under volted. Intel sets default vcore because that is where it's needed to be. E6400 should be able to run 3.2GHz with under 1.4V. Mine does 3.2GHz dual Prime95 stable with 1.385v.

[Ryzalis]

Well that didn't work. I bumped up the northbridge voltage to 1.45v (stock voltage is 1.2v, I think), and I'm still getting rounding errors in SuperPI at this voltage.

Any more suggestions? And a question, what is the FSB termination voltage and when would I need to raise it?

[Ryzalis]

Well, reverting back to 0502 BIOS (from 0507) fixed it for me. No multiplier adjustment and fewer voltage options though. :( I'm staying with this BIOS until ASUS comes out with a better one.

[heydo6]

funny how older bois fixed it

[AndrewZorn]

I tried that already (see second post). I still get rounding errors.
no, you didnt.
Still looking for an answer. I tried raising vcore to stock, but I'm still getting rounding errors in SuperPi. It's definitely not the CPU.
perhaps try... overvolting the CPU? i mean, it IS an overclock... see if like 1.35v makes no rounding errors
usually when you overclock you need more than stock voltage
its just how it goes