if i were you, i'll go switch the PSU to a better one, say a nice, strong single +12 rail ...

if i were you, i'll go switch the PSU to a better one, say a nice, strong single +12 rail ...

ok, anyway... usually, when i overclock, i'll find the highest my CPU can go with stock memory clock, in order to do that, you have to use memory divider, limiting your memory clock to DDR400 (200MHz x 2) or below to ensure, the RAM clock woudn't be too high), from there, increase your CPU FSB by 5MHz or 10MHz until you can't log on to windows, until then... just keep increasing , say at 250MHz x 9 (your CPU's multiplier is @ 9.0x), the system locked up, don't be afraid, reboot (hard-reboot if you have to) and decrease to say... 240MHz and reboot... should be ok to log on to windows, kindly run a short test, in this "section", i prefer use SuperPI 1M calculation, quick and simple, if your system shows an error in 1M calculation meaning your system's not stable, either play with vCORE or decreasing the CPU FSB until you managed to get passed in 1M SuperPI calculation... say, you passed on 235MHz @SuperPI 1M calculation, it's time for your system running on long run test, 32M SuperPI should be ok, but i prefer Prime95, run it overnight >5Hrs considered stable enough (for me la...)

same goes to memory, finding highest memory clock... set the CPU multiplier down to 5x, with stock memory clock (not using the memory divider), slightly increase the CPU FSB until you managed to find a good stable memory clock... , testing and testing all same as CPU test (above)...

and which one is good with min price

the EP430 you're using is good (not great) and already min price (RM100+ only) .
so far no problem for my stock voltage overclock 295 x 9 daily use as long as your casing doesn't comes with multiple drives and fancy fans and lightnings .

what proc and stepping you are use.. mine is 3000+ e6 and can't oc over 250fsb

---

Added on August 15, 2007, 3:56 ambtw i want to ask some1 know this.. a amd proc with "NOT FOR SALE" print on their face

my 3000+ was E6 too (sold to a close friend)
ADA3000DAA4BW
NBBWE 0613BAD

try add a little 0.1V for RAM and lower your HTT multiplier to 3X (over 250FSB)
lock CPU multiplier to 9X , disable the Cool 'n Quite
use the lowest possible FSB:RAM divider (100MHz / DDR200) and push slowly

--------------------------------------------------
NOT FOR SALE means not for sale

i have brisbane 3600+ and msi k9nu neov, using raptoxx 550w.. why in bios i can just set vcore until 1.3v?  because of the mobo limitation??
suggestion??

i orthos 2 mins got error edi.
FSB Bus = 235
LDT/FSB = X4
Multiplier = X10
Vcore = 1.375

running no divider (1:1) as my ram suppose to hold up to 250Mhz.
loosen clock to 3,5,5,10.
what settings should i use to check my OC in orthos?
noob here.

i think it's my Vcore. prior reading this, i increased my Vcore to 1.4. running orthos for 1 hours plus no problem. haha...anyways thanks mate. go to play more with it soon...

just wondering wat other settings should we use in the Orthos? i only use smal FFTs stress CPU as instructed by Kmarc.

glad to hear that , don't play with orthos , play in the BIOS, increase your FSB to 250MHz onwards , your Toledo begging you for more

Why not orthos? I usually use orthos to test for stability. I usually use the small FFTs to stress the CPU. However, some forumers actually recommend the following settings in orthos for stressing the CPU:

Custom setting:
Min FFT - 8k
Max FFT - 512k
Time to run each FFT size : 1 minute
Check the "Run FFTs in-place"

In addition, some would use OCCT for stress test.....

Why not orthos? I usually use orthos to test for stability. I usually use the small FFTs to stress the CPU. However, some forumers actually recommend the following settings in orthos for stressing the CPU:

Custom setting:
Min FFT - 8k
Max FFT - 512k
Time to run each FFT size : 1 minute
Check the "Run FFTs in-place"

In addition, some would use OCCT for stress test.....

I read from xs, some pple there say use blend test for orthos, since for AMD the mem. controller is built in. Blend test would be a better indication of stability.
In my case, usually if the first hour doesnt have any errors, then 90% the pc is stable.