Bogeyone001, there is more than one way to overclocking heaven.
I'm from the "slowly but surely" school; an overclocking exercise typically takes weeks for me (given my schedule and my penchant to do things methodically and systematically). I favor a lot of testing, especially when it comes to establishing stability. I believe in the "divide and conquer" approach, i.e., I break the overclocking process into three discreet parts, testing for the limits in the motherboard, the CPU, and the RAM; when I've found the limits for all three, I then proceed to integrate the limits of all three into one coherent overclock (with the aid of Gogar's AMD Overclocking Optimizer
). Stability testing for me is a very painstaking, methodical process as well, and typically will take around 40% of my time.
Other people go for broke right away, setting their motherboard's Reference Clock (aka HTT Clock or "FSB" in some BIOSes) to a nominal figure, their CPU's multiplier to the max possible (which depends on each chip; yours is 10x), adjusting the motherboard's HT Link Multiplier (usually to 3x), and putting the RAM on the biggest divider (to slow it down as much as possible). They also peg the VCore to some nominal figure, and then they test for stability, tweaking mainly their voltages as they go.
I think there is an inherent advantage in the slow-but-sure approach, and that is it's far easier to diagnose problems as and when they come up. By splitting up the process into separate pieces, you take most of the guesswork out of problem diagnosis, analysis, and troubleshooting.
But it's an obviously more painstaking way of doing things, so I can easily understand why it's not much in favor for the majority of people.
@ Blitz: Gogar IS on the first post of the thread.