Test setup & Methodology :
Piotke's Test Setup |
CPU | AMD Athlon 64 3000+ "Venice" |
Mainboard | Asus A8N SLI |
Memory | 2*512 Mb DDR600 Adata |
VGA | Asus 6600 GT 128MbAsus X800XT 256 MbGigabyte 6600 256 Mb |
The CPU was running at ~2.6 Ghz (9*290 Mhz, with the memory in 1:1 (synchronous)) I had some power under the hood. I installed windows two times; once for the NVIDIA based VGA cards, and once for the ATI card.
The installed software was:
Windows XP Professional
Service pack 2
Direct X 9.0c
nForce4 NVIDIA Drivers, version 6.66
ATI Catalyst Drivers, version 5.8 / NVIDIA VGA drivers, version 77.77
The applications I used to test were:
3D mark 05
3D mark 03
3D mark 01 Se Pro
UT 2004
Call Of Duty
If it was possible, I ran tests in both 1024*768 and 1280*1024 resolution. This was made easy thanks to a very handy tool: benchemall.
We have a mix of games and syntehectic tests for you, let’s start with the first one, 3D Mark 2005.
3D mark 2005
The first charts and I have almost a hit. The 6600 GT performs ~1,5 times as fast as the vanilla 6600. This is very close to the price difference. The X800XT doest the job "only" 2,5 times as fast.
3D mark 2003
In this case the 6600 GT is also 1.5 times faster. The X800XT shows its power in higher resolutions where it's 2,2 times as fast. When using a 1024*768 resolution, it's almost exactly double the performance.
And finally the oldest version of 3DMark that is still supported by Futuremark
3D mark 2001 Se
You can directly see that it's a rather old benchmark. Almost none of the newer rendering technologies is used, this causes the performance differences to be only factor 1,3 and 1,6.
Time for the games ->