All of the above responses are reasons that people go with DDR2-800.
1) It's readily available, and generall only slightly more expensive than lower speeds, if it all.
2) It's useful for the overclocking headroom, especially for lower-multiplier processors. (But even for higher multipliers.)
3) I hadn't heard about quads making greater use of available bandwidth. If the processor runs on a 1066 frontside bus, meaning it makes 1066 transfers per second, I don't see how the number of cores has anything to do with that. But I'm no expert in that area.
And just a side note, RAM running on a 266 MHz frontside bus is running at DDR2-533, not 633.