fsb 1:2 is that good?

I oc my e8400 to 3.6 (fsb 400)
and my mem went to ddr3-1600 (fsb 800) (well it's actually dd3-1600)

I always thought that fsb 1:1 was best, so should I change something?

Well, the multiplier is at his max (9)

No.

Higher speeds (MHz) mean more bandwidth, which means faster read/writes and the like. It might not be tangible in most situations but it is there.

CAS 7 DDR3-1600 = 0.0043
CAS 5 DDR2-1066 = 0.0046
CAS 4 DDR2-800 = 0.005

Lower is better, and I just did time / MHz because I don't feel like dealing with nanoseconds by using MegaHz instead of just Hz. The idea is the same, just add zeroes.

Running in 1:1 is the most "efficient" as the story goes. In all reality the higher you can run the bandwidth the better the performance will be (synthetic). You just may not see a difference in real world benchmarks.

Well eventually there is a real world difference.

We don't run DDR400 in our machines anymore (well new ones that is).

The real world differences between DDR2-800 and DDR2-1066 might be negligible, but once you start talking DDR3-1600 and the rest, it might make a bit of difference (especially in memory intensive applications.. like Super Pi).

I wish I could hit DDR2-1400 MHz.

Um.. yea.

It'll be DDR4 by then though.. Sandy Bridge.