eh it sort of is, sort of isn't. 4:5 is definitely harder on the northbridge for p965. mbench for sure is just a synthetic, and realworld isn't that represented. I mean, there's not much difference between 1:1 and 4:5, although you'd probably see a little improvement in superpi, but then again, what's that mean in relation to anything real, I suppose.

But numbers still mean something. 4:5 shouldn't be any worse than 1:1. Any problems with running async with the upscalar should be compensated by the better latency/bandwidth.

I'd say, give it a go and see if you notice any difference.