Interesting. I was always curious just now much cache mattered in superpi--there are tons of people that make it seem like it's such a big difference (like between a E6300 and 6600), though there was no proof. Besides, I'm wondering if 0.3s can just be attributed to randomness, or different background processes, and not necessarily due to cache size. Though if it's always faster for the 6700, then I guess it is the cache--just didn't realize how little a difference it makes.
Yeah, my brother and I used the same 'profile' for our running processes, to make sure we had a semi-stable environment. We even set our RAM to the same timings and voltage for each speed test. He has the copy of the results, but they went something like this;
2.6ghz = 18.8sec average for e6700 vs. 19.3sec average for e6400
3.2ghz = 16.7sec average for e6700 vs. 17.0sec average for e6400
3.6ghz = 15.1sec average for e6700 vs. 15.6sec average for e6400
and so on. Our tests stopped just after 3.7ghz, as my brothers rig is air cooled, and couldn't hang w/ the temps on 1.58vcore!
OOhh well, anyways, you get the idea.