the problem with benchmark tests is that they are predictions of how processors will handle real-world applications... and they're not necessarily accurate predictions. a chip can score well on a benchmark test that's designed to use certain things the chip does well... and the same chip can still completely suck at running a whole variety of popular programs.
the best you can do is run the processor through a sample of bench tests and popular programs, and get a rough estimate of how it performs.
Bookmarks