First I'd use a low multiplier for the cpu, like 9, and see what the memory will do. I would give the memory 2.9V and test it at Cas 2-2-2-6 to see how high it goes stable. Then benchmark with Sandra, Aida32, UT2003 and 3DMark01. Then back off to more conservative timings, 2-3-3-7 or 2.5-3-3-7 and see how high that will go and do the same benchmarks.
When you've found optimal FSB and timings for the memory start raising the multi.
I test stability with memtest86 and prime95.
When you've found optimal FSB and timings for the memory start raising the multi.
I test stability with memtest86 and prime95.