Ya know, I get that these guys' day jobs aren't to be actors or showmen of any type, and I get that the point they're making is both valid and factual because pretty much everyone, including Intel users, know that sysmark is biased, but MAN, listening to these guys talk was like listening to some kids read aloud in my 8th grade English class. One of the most important things when trying to deliver a convincing argument is some degree of passion and personality when arguing your point. Having facts is, of course, great too, but without showmanship, no one cares as much as the flashy guy spouting ignorance, which is the big problem facing certain political parties in our country right now.
Criticisms aside, the best benchmarks you can find are probably real world examples of the components you're considering being paired with the tasks you're actually running. As much as I love Unengine's Heaven, when considering my purchase, I used youtube as well and found a fantastic video showing gameplay and framerates on various games where the only difference was the CPU, graphing the framerates for each throughout the course of the entire gameplay as it progressed. These results, incidentally, were pretty much in line with the 6% or so delta the video claims to be more accurate, with the 8350 keeping up pretty well with the i5. It's easy to say "just go with the faster chip, whatever" until you look at pricetags, and then you realize that having $50 more to put into your video card goes a lot farther to aid your performance than getting the i5. The fact that I also knew that I could overclock the 8350 really well pretty much sealed the deal and made it my choice. I was so pleased with its performance and value, it was going to be an auto-include on my other build, but I had to get an 8370 instead, since the 8350 was out of stock when I made the purchase.
Well I think that you need to use more than one benchmark to get somewhat accurate results. Not to mention that real world performance would probably be slightly different than synthetic benchmark results. Anyway it's good to know some background information about the the benchmark software.
Retrieving data ...