A Deja Vu Over-Clocking issue

[BonzaiDuck]

First -- I completely agree that my rig is at least a generation old. It's Intel. It's a single-core Prescott. But the board is a good over-clocker, and the memory modules are rated at DDR500 with very competitive latencies -- for those speeds anyway.

It's been a great video processing system, and I occasionally play some games with it. And until the income from my rental property has resumed flowing again, I'll be hard-pressed to shell out for an SLI-based system -- or the Athlon 64 X2 it will use.

Meanwhile --

I was experimenting with over-clock settings, VCORE settings and MEMTEST86. I had been too cautious previously in setting the VCORE, and have now discovered that I can run the VCORE up to as much as 1.42V on my P4P800 SE mobo without much risk to the processor.

So I decided to disable the ASUS "Lock-Free" mobo feature, over-clock the FSB without underclocking the CPU multiplier, and found a nice, stable setting of about 3.74 Ghz (with a 3.2E Prescott) where the memory runs at DDR468, or an FSB of 936 Mhz. I figured that the L2 cache-size of 1MB would probably compensate for a slightly slower memory setting.

I ran MEMTEST86 all freakin' night without errors. Something like 20 iterations on the tests, and with the VCORE set only to about 1.4125V .

So I try to boot into Windows. Error messages post about the Winlogon.Exe program failing to recognize hardware -- unable to validate activation -- and the system either crashes or makes it impossible to logon.

I tried dropping the external frequency to 230, lowering the CPU speed to around 3.6 Ghz, and the system actually boots, but my AV software reports a failure to load some DLLs, and refers -- again -- to "an activation" problem.

Anyone have any suggestions? I KNOW I've seen people discussing this issue before . . . . .

 

[trexpesto]

what is your chipset MHz rating, and voltage

 

[BonzaiDuck]

yo, trexpesto !! --

Obviously, the i865PE chipset is only "rated" for 800Mhz FSB, but the P4P800 SE is a good over-clocker.

I had previously followed OCZ's recommendation for using their DDR500's with the P4P800 boards, and cautiously set my AGP/VDDQ -- what I think you mean be chipset voltage -- to 1.6V. Some people set it to 1.7, but I'm a cautious guy.

Since my last post to this thread, I've discovered a factor which completely eliminates the XP boot-up problem at the higher CPU speeds -- 3.7 Ghz and above.

It is the doggone VCORE !! After poking around various forums and (archival) posts (because this rig is surely becoming obsolete with AMD Ath64 X2, Intel Smithfield and Pressler) -- I discovered that most people using this CPU (3.2E socket 478) set the VCORE above 1.4V. Mine had been set at 1.3875V for the entire six months I had been running this sucker's FSB at (external frequency) 250 and 1000 FSB. How it passed the stress tests -- and I ran plenty -- I can't fathom, but just about every indication I have from the web-searches shows people pushing the VCORE to at least 1.42+V-something.

It may be that my PSU lost a little of its adjustment -- it's a PowerStream. I don't know. I tweaked it just a tad.

All I know is that I pushed the VCORE setting to about 1.43V, and I'm running a series of OC tests. It is not running at DDR500, but only at DDR488 at the moment, but everything seems back to normal.

It seems to pass PRIME95, and S&M. I have another question for which someone may know the answer regarding S&M -- I'll post as different topic.

 

[BonzaiDuck]

I figured out my problem here, and I really feel stupid. for instance, why would I have to adjust the VCORE upward for stability after it had run stable for many months at near-default levels?

Anyone interested -- I'll explain.

Is anyone using a DDR Booster in their system?