The AMD enthusiasts tend to experiment more with divider settings that depart from 1:1.
I follow some assumptions that were part of my curriculum during a 1990's computer architecture and organization course. The machine is a pyramid that has high-speed, low-volume, high-cost storage at the top (CPU registers), and relatively low-speed, high-volume, less-expensive storage at the bottom (hard drives and optical). Buffers and caches open bottlenecks between fast and slow levels. Hence, L2 cache. If your memory is fast to begin with, you should need less L2 cache to bridge the gap, but "more is better", anyway. This is one reason I am carefully moving toward the idea of a Prescott with C0 stepping and 1 Mb of L2 for over-clocking, as opposed to a Northwood with only half the L2. since I was able to run the memory at the flat-out spec of DDR500 with a 2.4C Northwood, but I can only run it now (in a practical sense) with a 3.0C at DDR480, I'm thinking the processor is my limitation, and my mobo's ability to unlock the Prescott C0's multiple should give me even greater flexibility.
With this in mind, I focused my money on GUARANTEED memory speed, hoping for a stable 1:1 configuration, so I bought some insurance with EL "Gold" DDR500 as opposed to taking my chances with the Platinum rev. 2 OCZ's.
It is nice to have low latencies, like 2, 2, 2, 5, but myths have been dispelled about the importance of CAS, and OCZ has implemented some features that increase bandwidth anyway with looser latency settings. Further, slower speeds and tighter latencies yield proportionately less gains in performance than slightly looser latencies and higher speeds, so again, I didn't plan on running my memories as DDR400 or even DDR433 -- hoping for something between DDR480 and DDR533. SiSoft Sandra proves it: I have an "edge" with DDR480 over DDR400 CAS 2.0 -- just not a "proportional" edge matching the proportional increase of system-bus speed in Mhz.
You have a range of viable options, and choosing one versus the other leaves you with more or less flexibility to do something else. I am the last person to say "throw money at it and it will be better." But in order to have optimum flexibility, you cannot always put the price-factor at the top of your priority "pyramid" (ha).