New results...nothing special

[Spiegel]

Well, an idea stuck me, so I tried it...but it didn't work out too well.

I was running 200x9=1800 with the RAM at 2-2-2-5, as seen here:


and here



Then, I tried to see how high I could get at x12....I got to 155 w/ RAM still at 2-2-2-5 and got these results:


and




Obviously, my scores went down...memory bandwidth especially. So the moral: my settings are back where they were, and I've finally decided to leave OCing alone until I get my new PSU.




On a side note, can anyone tell me what I'm doing wrong? Generally, I loosen my RAM timings up, then I start trying to OC. Thing is, I never get an "unstable OC" and then am able to go back and up the voltage or whatever. Rather than being unstable, my machine just refuses to post. I go from stable and fine to BEEP BEEP...."Your CPU may have a problem....Your CPU may have a problem...." It'd be much more convenient if when I screw up or go too far with an OC, I can go back and fix it, rather than have to wipe the CMOS every time...any suggestions?
_________________

[Google]

[meobius]

of course the bandwidth would go down
the 9(12) is the CPU multiplier, where as the 200(155) is the system bus, the performance of a system is usually more affected by a change in bus speed than anything else. When overclocking you always want to boost the FSB(200) as high as possible. Try dropping the multiplier (9) down and bringing the FSB(200) up, you'll probably have to relax the timings on the ram
_________________
If time is of the essence,
than why is it that we waste,
that which holds such great importance,
spent without a trace.