Hate to burst your bubble DragonCommando, but benchmarking is the only real way to compare the performance of two cards, you have them all run exactly the same test and compare the results. You can't compare clock speeds and automatically know which card is better, the true performance of the card is the combination of all the specs. But we don't know how they are combined, so we cannot tell from the core speed + memory speed + amount of memory + all the other stuff in there how one card compares to another card. And sometimes you cannot compare specs directly, like ATI and Nvidia have different ways of counting shaders, so an ATI 4870 has 800 shaders and a Geforce 280 GTX has 240 shaders (or something, dunno exactly). The 4870 also has a faster core clock speed than the 280 GTX. Does that make the 4870 faster? No, it doesn't. In fact it's slower. But how do I know that? After all the specs tell me the Radeon should be faster, right? Benchmarks tell the real story, even if they don't match exactly from site to site, they all follow a similar pattern.