4/732 = 5/915 = 1/183 ms = Timings are identical
If you have AMD, then it's 915 all the way.
On Intel, it should still be 915, latencies are the same effective values, you still have extra memory bandwidth for when the fsb doesn't hinder you. Sometimes intel systems seem to react negatively to running memory multipliers though, but you need to benchmark to see what happens.
I'm still not entirely sure why the extra bandwidth isn't advantageous to DMA....