n00besque O/C woes: please help!

[Wigwam]

My system spec is as per sig [pertinent details: 2100+ t/b B; A7N8X-del rev2; twinmos pc3200 RAM running 6-2-2-2; 9500pro; maxtor DM9+ 120gb]
the machine has been running fine for 9 months on fsb 166 X 12 multi; i played a little and upped it to 180 x 11.5 = 2070 and it ran fine for a day [then i decided to return to 2400+ speeds cos i was doing research and didnt wanna risk it going kaput.
all this on stock vcore.

Today i changed it to 200 x 10 [again stock vcore] and it booted fine; feeling adventurous i upped it to 10.5x200 [but reduced the ram timings to 8-3-3-2.5] and it just wouldnt post - cdrw/dvd flash and on the 1st occasion after a cold start it displays the cpu speed at 2100 MHz but then complete freeze; i occasionally get screen corruption with flashing horizontal lines; if i press reboot then i get no video signal and it still freezes before the "ticking" sound of the ram being checked.

I dont understand what would be the likely cause for this problem - could it really be that sensitive to not working at posting 2100 when it ran fine at 2070?
also, as a genral principal is it at all possible to determine why an O/C failes to post just based on what it does [eg in this case stopping just before ram check]? i dont see that is can be ram problem cos the fsb was not changed between 2000 and 2100 and it was okay at 10xmulti with tighter timings.
could it be the video [but why should that be cos lots of folks here said it was agp/pci locked on this board: it is set to "auto" in the bios]
finally, could it simply be that the diff between 2070 and 2100 was enough for it to need more vcore?

i would appreciate help and insight in this matter

ps tip: when clearing the cmos and restarting it is always a good idea to reinsert the cmos battery; i didnt at first and wondered why i was getting all these weird glitches!!!!

 

[Revo]

Well dude being a newbie myself I had a similar problem and increasing the Vcore voltage solved the problem for me. A slight increase in CPU frequency from my experience can lead to the problems you have described. Moreover, each CPU is different and therefore people have varying OC'ing experiences so, when I suggest an increase in Vcore it may or may not work for your particular situation. Lastly, you should know that a large increase in Vcore could be quite detrimental as I found out when my power supply died on me! In any event, I would suggest that you clear the CMOS and increase the Vcore conservatively until your problem is solved.

 

[myocardia]

Revo's right. Put your vcore to 1.75v. That's a safe voltage to leave it at for a long time. Also, I would suggest using this: link to ClockGen. It's a Windows overclocker. After you raise your vcore a little, just boot to Windows, open ClockGen, and raise your fsb to whatever you want to try. If your proc. can't do it, Windows will freeze, and all you have to do is hit the reset button, wait for Scandisk to run, and Windows will reboot at the same frequency you started at (for instance, 10x200 or 10.5x180). You'll need to download the file called CG-NVNF2, the file for all nForce2 boards. I use it, and I love it.

 

[DeRailur]

Today i changed it to 200 x 10 [again stock vcore] and it booted fine; feeling adventurous i upped it to 10.5x200 [but reduced the ram timings to 8-3-3-2.5]

Correct me if I'm wrong but you shouldn't have to loosen up the RAM timings if you are making a multiplier adjustment only if you want to raise FSB.

 

[Wigwam]

Originally posted by: DeRailur

Today i changed it to 200 x 10 [again stock vcore] and it booted fine; feeling adventurous i upped it to 10.5x200 [but reduced the ram timings to 8-3-3-2.5]

Correct me if I'm wrong but you shouldn't have to loosen up the RAM timings if you are making a multiplier adjustment only if you want to raise FSB.

yeah i know - it was a somewhat belts and braces approoch but it stilll didnt help

 

[Pilsnerpete]

nVidia also makes an oc'ing utility. I have it someplace here...ya, it's the "nVidia system utility". Look for it on their website.

I also have a T-bred B. In fact, you and I have the same system except for the hdd and vid card! I have the cpu @ 1.825v in BIOS. 2.19ghz in system properties/199 fsb. Been running it this way for months.

 

[Wigwam]

have you used the nvidia utility? i was considering it but posted here for views of ppl who have tried it and pretty much got unanimous "it locks up my machine" replies so steered clear for the time being.

what HSF you using pilsnerpeter? i only got a spire falconrock II and while it is nice and quiet and better than stock, still i am aware upping the vcore might my raise the temps too much which is why i am trying to get the best from stock at the moment.

it is a dilema though of running 12.3 or 13 x 166 [which it does manage on the stock vcore: which is even wierder cos thaat is higher than 2100!] or running 200fsb at 10 x?!?!?!

i didnt realise upping the fsb might need more vcore rathr than just the overall speed?

 

[myocardia]

Wigwam, you have reached the fsb limits of your proc., that's all. You can either leave it where you have it, or raise the vcore, and hope for stability at a slightly higher fsb. Most processors, though, don't go considerably higher with huge vcore increases than they do at stock voltage.

 

[Wigwam]

Originally posted by: myocardia
Wigwam, you have reached the fsb limits of your proc., that's all. You can either leave it where you have it, or raise the vcore, and hope for stability at a slightly higher fsb. Most processors, though, don't go considerably higher with huge vcore increases than they do at stock voltage.

myocardia, not sure if i am missign something but i am not sure it is as simple as that:
it posts/runs [albeit somewhat hot!] at 13x166 = 2158 so raw speed isnt the wall
it also runs fine at 10x200 = 2000 so fsb is okay here [i mean it is an AUIHB overclocking champ!]
so why wont it post at 10.5 x 200 = 2100?????

i take your point about the benefits of big vcore increases though and woud stay clear of those; i seem to recaall that when cpus are dying you need ever-increasing vcore to support them.....

still very mystified!

 

[Wigwam]

bump for more info

sadly it wasnt as stable at 10x200 as i thought as it locked-up in the sandra burn-in test so back at 12x166 [12.5x166 runs prime95 fine but a bit too toasty for my wee hsf]

i dont get it: a 200fsb-rated board; an unlocked o/c-king chip; PC3200 ram and it isnt even stable at 2.0GHz on the 200fsb but okay at 2.08GHz on a lower fsb...

i have read on the nforcershq that sometimes the 10 and 10.5 multis are flakey on this board so maybe that is the problem but i hope not!!!

 

[Wigwam]

bumo for more help please
i got a 7F [0D] stop error when testing at 9.5x200 [with pc3200 mem at 8-3-3-2.5] all at stock vcore and vdimm
i will run memtest 2nite but it has been fine for 9months at 166fsb....

i still dont see why i am having these troubles!