Originally posted by: chizow
Originally posted by: dpopiz
well I'm having terrible luck here...maybe you guys can help me out:
- first of all, everything's completely stable at normal speeds (133x13), 133MHz DDR
- then I upped the fsb to 166 for 166x13=2150, ~42C and things became a little unstable, and it hung when I tried to run pcmark (ram was set to "auto" in the bios, so I don't know if that means it was at 133 or 166
- anyway, then I decided to try setting the ram manually to 166...that caused it to hang every time on entering the bios or booting and required me to use the "safe mode" jumper to get the fsb and ram back to 100 temporarely
why am I having so many problems? why aren't I getting "2.3 no sweat!"
It can be a lot of things, if your rig in your sig is your current set-up, the first red flag I see is the chipset. You're probably op'ing off a 1/4 PCI and 1/2 PCI divider, so upping the FSB to 166mhz alone is a stress point.
What speed is your Kingmax RAM ram rated? Do you know the speed of the RAM chips themselves in ms? That's another potential bottleneck.
How about your PSU? If you have a voltage/temp monitor recording program, run it while doing benchmarks and see what your rails and temps are doing.
Temps seem OK, but again, I think that board takes its temp readings from underneath the socket which translates to inaccurate actual temps. Also if you are getting a reading from underneath the socket, fast changes in CPU temp will be masked b/c the change in temperature isn't real-time. So you might see a completely normal temp reading right before you crash, but you'll never see the true temp b/c of the lag in heat build-up inherent in socket thermistors.
I'm thinking you might want to access the higher multipliers with that set-up and go for a moderate FSB OC and try to push your multiplier instead. You won't be stressing your RAM and your PCI bus as heavily that way.
Chiz