Yep, since its inverse proportional, you'll always need more and more MHz for the same absolute time difference. I don't think there's any way to have such a formula. We can compare efficiencies at a given set multiplier/ratio/timings, but once those variables are changed, the efficiency will as well. When I was doing those runs I usually eyeballed efficiencies at the next multiplier step by adding the difference in performance product between the previous two, but that turns out to be too optimistic and is better for nothing but a rough rough estimate.
You guys are all making me damned jealous, nice to see all this action going on. Love seeing 3ho6 right at the 59.984 mark for 12.xx--that's the sign of a totally squeezed setup.
I wish I had the time to do 32m runs on P35, you'd crap if you saw the kind of mem speeds I was seeing.
DDR3 is def where its at now tho...keep defending DDR2's honor nonetheless dinos.
Originally Posted by Gautam
Yep, since its inverse proportional, you'll always need more and more MHz for the same absolute time difference. I don't think there's any way to have such a formula. We can compare efficiencies at a given set multiplier/ratio/timings, but once those variables are changed, the efficiency will as well. When I was doing those runs I usually eyeballed efficiencies at the next multiplier step by adding the difference in performance product between the previous two, but that turns out to be too optimistic and is better for nothing but a rough rough estimate.
Reply With Quote
Bookmarks