Futuremark 3D MarkThese synthetic systems tests from Futuremark give you a good overall idea of performance between the video cards: you should not solely rely on these results though as the game benchmarks do not always reflect these rankings.
I always start with testing 3D Mark performance, it was already then that I noticed how the Fatal1ty 8500 GT was not really an extraordinary overclocked video card, this time XFX behaved polite which resulted in only few 3D Marks difference between XFX and NVIDIA reference clocked samples. Overclocking the video card brought us to 32% higher performance in older 3D Mark benchmarks while 3D Mark 2006 resulted in a 43% higher benchmark score, now that's some Fatal1ty!
TemperatureConsidering the low power consumption like we saw with our
GeForce 8400GS round-up we think there shouldn't be many problems keeping the GPU at reasonable heat level. We ran 3D Mark in loop and logged temperature with the latest version of Rivatuner, here are the results:
Even without using active VGA cooling we managed to get a healthy overclock on our video card. The temperatures went high but nothing to really worry about. Our housing does have decent airflow and we do want to say that you may not get the same overclocked as we did, not only will sample variance come in to play, the GPU temperature may be 10°C higher because of reduced fresh air intakes and that will also have some influence on your video card overclock. Don't let this scare you away though, the default clocked XFX Fatal1ty is no where near overheating levels and will provide deadly silent 3D video processing power without any issues temperature wise.