Memory Question on my stability

I am prime95 stable using 2 instances of Large FFT. 1 Instance at realtime priority and the other set to High so it doesnt lockup the pc.

At 233mhz FSB it reboots after a few minutes doing the exact same test. At 232mhz it locks up when it reaches 12 hours into the same test. Is this a ram issue being that its at its 1:1 limit or a NB heat issue? I have the AGP voltage set to stock which also controls the NB voltage. Could this also be a case and if so wouldnt this hurt the video card if I increase its voltage?

------------------------------------------------------------------------------------------

I just now did a little playing around with the AGP voltage. I set the FSB to 240 and the agp to 1.6 and it wont post. When I set it to 1.7 it goes to the windows logo and blue screens afterwards. I then reduced it back to 1.5 since that obviously wasnt stable.

any sidjestions?

Have you tried memtest86?

i am 24hour stable doing Blend at realtime priority with nothing in the back besides the OS. I can achieve this stability doing this test at even 233FSB. Once I start doing Large FFT stressing I have to lower it to 231 or it crashes.

u may have hit ur limit, but try this. bump the ddr voltage to max, bump the agp to 1.6-1.7 and try bumping the vcore to 1.65v( with asus boards the nb voltage is tied to the vcore so when u raise the vcore the nb voltage also goes up). also try to turn off ur performanse acceleration related mam (or turn the setting down). i have found that hi latency ram doesnt do well with it at hi fsb.

In 3:2 divider of course correct?

if u havent already tested the max of ur cpu then u want to do that first with a 3/2. however if u already have done this and want to test the memory then run it at 1-1 (i am assuming from ur sig that u are testing corsair 3700)

I reverted the bios back to an older version which brought me to partial Pat. Which in turn allowed me to once against use the memory dividers since the Full Pat hack disables them. Now as I experimented with just using the CPU stressing I was able to boot into 3.75 and such but froze all the way to the mid 3.6's. At low 3.6 i was able to get into windows which was nice. I started stressing doing the same tests as when I had the full bios hack. 238FSB was able to stress for the longest time till it froze after 30 or so minutes. Anything higher then 238FSB was the same. I also had to put the vcore to 1.7v in the bios which was bad so if i have to do that then im not doing it unless i do the vcore mod which is a hastle.

Anything below 238FSB which is going farther away from 3.6, my origional goal, just isnt worth it. Im perfectly stable at 231 while in the Full Pat mode. Sucks majorly but this is my stopping grounds. grrrrrrrrrr.

Im sure i can reach a higher stable area by increasing the vcore which it did show while stressing in the high fsb areas but its not worth it nor the hastle to do the mod.