Issues Overclocking a 3500+

[biff420]

Alright I am starting to get a little frustrated. I have tried several overclocking scenarios but none seem to work. The problem is usually the motherboard itself. I can get a rock hard stable overclock configuration to work in windows with Prime95 but as soon as I reboot the computer will either not post at all, complain about a bios error, or boot without any video. Everytime this happens I have to play around with the CMOS reset prongs for about 15 minutes before I can bring it back to life with a full video boot. By then, of course, I have to start all over with messing with the settings. Its like giving CPR to an ant named Shiavo.

The best stable overclock for windows/prime95 sake has been to set the HT to 4x, the mem divider to 166, FSB to 240, and the multiplier to 11x. This results in a speed of 2641 Mhz. I don't really mess with the voltage at all becuase when I load up Motherboard Monitor it stats the voltage to be at 1.65V even if I set it to Normal (1.4V) in the bios. It seems to be automatically adjusting for the increased power needs. Also the memory should be at a 199.2 FSB with the memory divider so it should require no change in voltage. And by setting the HT to 4x with that FSB I end up with a HTT of 960 which is under 1000 and should be acceptable.

My system config is:

Gigabyte GA-K8NXP-SLI
AMD Athlon 64 3500+ Winnie (90nm - 2.2GHz stock)
Corsair XMS TWINX1024-3200LLPT 2-3-2-6 1T
2 x eVGA 6800GT (SLI)
2 x 80GB SATA (Raid 0)
1 x 120BG IDE
The rest is the usual crap.

Please post any suggestions on what to do??

Edit: Temps for 2640MHz are about 34 idle and 52 full load.

 

[biff420]

Bumb...c'mon guys, I'm really suffering here.

 

[Diogenes2]

Sounds like you have just reached the limit of the chip...

2641 Is a decent P95-stable overclock...



What reason do you have to believe it should go higher..

 

[biff420]

It not a matter of going higher. I am happy with the 2641 OC. Its a matter of being completely stable in windows but having the bios errors. I would think that if Windows is happy with it then the bios should be more than happy with it.

I need to know if there is something wrong with my board or if anyone else has seen this type of behavior. I am emailing Gigabyte to see what they can say.

 

[Diogenes2]

Ooops.. Sorry I missed that..

Have you tried flashing the Bios ?

I agree it sounds like a MOBO problem...

 

[zizo]

Fix the vcore first. 1.65 is too high if you are on air.

 

[biff420]

Like I said. The vcore is set to normal which is 1.4v. It is changing it automatically. Maybe if I put it at 1.4 instead of normal as the setting. I will try that. Do you really think that is causing these errors?

 

[zizo]

What's the vcore in the bios system monitor? try speedfan to check vcore in windows.

 

[Shimmishim]

it shouldn't jump from 1.4 in bios to 1.65 in windows...

make sure you set it to 1.4 or 1.65 in the bios and not on auto.

 

[biff420]

OK sounds like a plan. Do you really think that this is causing all the problems with the bios boot. I would think that overvoltage would only cause error once in windows and the processor is being taxed.

 

[imported_wyrmrider]

good case cooling, ram cooling, chipset cooling
ps ok
after dbl checking the above
set 3x HT does that board support 1000 HT or 800???
up the ram divider and test again- to eliminate ram
then set cpu multiplier low and make sure the HTT is stable
let us know what you come up with

have you been to a good gigabyte forum?

 

[biff420]

Alright....the voltage setting seems to have fixed the issues but who know what tomorrow brings. I never would have thought that would have solved the mess I was getting from the bios.

My current setting for others reference are:

HT = 4x
Mem Div = 166
FSB = 236 (240 was unstable)
Mult = 11x
vcore = 1.55 (tried 1.4, 1.45, 1.5 all were unstable)
Everything else is at default

So thats clocks at 2.6 GHz quite well.

Now I get a 10100 3dmark05 score!