Originally Posted by SpacemanSpliff
Unless I'm not catching your drift here, I disagree... running clock for clock in as identical a set up as possible gives one a very direct baseline for comparison. If one wants to truly show and compare the potential of two CPUs, GPUs etc... then it makes the most sense to compare baseline performance as well as peak performance since unlike around this forum, not everyone cares to overclock and fully push their systems. And for those of us that do it provides a better metric of the full performance spectrum of a product and how well it fits into the plans one might have as far as OC potential, overall power consumption, folding/editing performance, gaming performance, etc.
Look at it this way;
when haswell-E comes out with DDR4 support should we demand that all CPU comparisons should be made with DDR4 as fast as the DDR3 used by the older setups? No of course not. Being able to support higher speed DDR4 will be a feature for haswell-e.
Similarly being able to use higher speed DDR3 is a feature for haswell. Its IMC is better than that of the 8350 and will handle much higher speeds. So take a situation where you wanted to see the max potential of the chips at stock. You would pair both with the fastest possible memory available of course, to maximize performance. However haswell would be the only one that could run it at rated speeds. Should you then cripple haswell just because the 8350 can't run memory as fast? Of course not.
Haswell can run faster memory than Vishera can. And because of that running Haswell with faster memory should be a norm. Unless you're doing some build comparison at specific budgets etc.