If you need a benchmark to tell you it is good then you just wasted money, you should see a difference. I don't play games, so my needs may not be the same as yours, but I want refresh rates of at least 100hz @1024*768 that redraw the windows instantly when switching between programs even when the machine is drained of resources. The ATI does not do that, even with 64mb of DDR. The Abit does.
Benchmarks are meaningless for real life computing where things like windows eye candy and stay resident programs that work at random will change performance and would not be tested in the benchmark. Sorta like overclocking, I remember an old post, everybody posted clock speed and seconds to open photoshop, time from click to open. The overclocked machines were no better, in fact, it sounded like a crapshoot actually.
Same with a more recent post where I put up a PHP script that loops echos and clocks the time. Everyone had drasticly different results that were not as expected considering computer specs and PHP installations (module/CGI, version, etc.) A JSP version of the same script made things more confusing, meaning the same operations will work faster with one technology on one machine and slower on others (the JSP was always faster or tied to the PHP.)
My point is, if you can't try it out, read the reviews and I would not rely too heavily on benchmarks. Rumours about benchmark companies being paid by various vendors to make the tests utilize strong points of the hardware don't help either...
If you end up with a CRT, make sure you don't mind the two black lines that appear on many of them, or buy one of the very few models that don't have the lines...