overclocking 3.2c problem help plz!

[mc2k]

Hi first of all here are my pc specs
P4 3.2c
asus p4p800se
kingston hyper x 2x512mb KHX3200/A ( refering to the site 2-3-2-6)
club 3d x800 pro 256 mb
Tagan 480W psu

couple days ago I decided to o/c my systeme. according to cpu-z : the ram timings are @ 200 fsb (SPD) 2.5-3-3-8 (instead of the 2-3-2-6) I did ignored it and started to o/c. I did with cpu/ram 1:1 and came to 228 fsb (229 wouldnt boot windows) according to cpu-z it was now 3.65 ghz and ram timings 2.5-4-4-8 (still SPD). I did lock the agp/cpi, and change the vdimm to 2.85V, and the vcore was set to auto(vcore isnt constantly under full stress its even below 1.55) well I tested first with 3dmark2001 35x per test (3.5 hours) didnt crash att all and the score was the same as I 1x. But today I tested with prime95. after 3 hours and 1 minutes the toture test stopped and reporting this:FATAL ERROR: Rounding was 0.49609375, expected less than 0.4(during the test max temp: according to asus probe: 48-49C)
Hardware failure detected, consult stress.txt file.
What must I do? changing vcore?(auto) or lower vdimm(2.85), change timing(tho its already high), or something else? plz help me

<P ID="edit"><FONT SIZE=-1><EM>Edited by mc2k on 01/08/05 01:24 PM.</EM></FONT></P>

 

[Scout]

My guess would be raise the vcore a bit... you're getting pretty fast for that Northwood. I'd also lower your vdimm some. Unless those memory modules really like that high voltage, you could be hurting stability more than helping it at 2.85

Scout
700 Mflops in SETI!

 

[mc2k]

what vcore should I set? and what vdimm?

and Does asus probe read the temperatures correctly( cuz I heared some rumors that asus mobos reading temperatures inaccurate about 5-7 degrees below. dont I know if it is the truth. and How hot may a northwood run?

 

[Scout]

I have to warn that any time you increase vcore, you are doing it at your own risk... That being said 1.6 volts is pretty safe. Crazy overclockers will go higher, but you risk shortening the life of your processor.

I'd put the vdimm back to 2.6 or 2.7 and see what works... obviously, you want it lower if you can achieve stability due to heat issues.

Temperature measurement can really vary... I don't know that Asus does any worse job than the other manufacturers. I have a Northwood that runs at about 50 C under load at 3.2 GHz with a stock cooler at 1.5 volts. (It's an overclocked 2.26 processor). If you're getting readings close to 60 under sustained 100% load, then I'd be concerned.

Scout
700 Mflops in SETI!

 

[mc2k]

yes I have put vcore to 1.5875(anyhow according to asus probe/cpu-z the vcore underload is still beneath 1.55(1.52-1.536, I thought I locked to 1.5875 tho), vdimm to 2.65 but @ 228fsb it wont boot , its now set to 225 fsb it booted, ram timings still the same, I am now running prime95 hope this time wont show error, btw how long must i run this proggram to be sure if it is rock stable?

 

[Scout]

It's pretty typical for the CPU voltage to sag lower than what you set it for in the BIOS, especially under load... I simply set the BIOS a but higher till I hit the voltage I want.

Not sure how long you need to run the Prime95 to guarantee stability though... I think the only way you can really assess stability is over a period of weeks; run the system under a variety of conditions under heavy loads and if you never have a crash you've got a stable overclock.



Scout
700 Mflops in SETI!