For a DX8 game benchmark UT2004 was used. The
resolution of 1024x768 and lower were not used because results are CPU-bound.
Maps used for benchmarking were dm-rankin, as-convoy, and br-colossus. Good
performance in any of the resolution/settings displayed.
DIRECTX 9 BENCHMARKS
For testing DX9 Aquamark 3 is used to represent the synthetic benchmark although
using an actual game engine. AM3 still provides a good test for evaluating
graphic card performance in a graphically stressed DX9 environment. Results were
very consistent and even exceeded previous scores I had obtained on nForce4 x8
SLI motherboards by as much as 6% using this same SLI configuration.
The benchmark results displayed solid numbers and
the K8N Diamond equipped with both the 6800 GS cards in SLI and the single 7900
GT exhibiting good performance with excellent stability. In this benchmark the
7900 GT came out convincingly ahead at all benchmarks and settings although
close at 1600x1200, 4xAA/8xAF.
For the DX9 game test I used the current popular game F.E.A.R.
Comparison scores were obtained using two x8 PCI-e motherboards (EVGA
133-K8-NF41 and an EPOX 9NPA+ SLI) benchmarking the game at a resolution of
1280x960 in the in-game settings with antialiasing and ansiotropic filtering
applied as identified. The variable, of course, being the different motherboards
with all other components remaining the same.
F.E.A.R. Test Results
The MSI K8N Diamond Plus maintained performance level much better as
antialiasing was applied and increased. While the game is playable on the other
two nForce4 x8 PCI-e SLI motherboards at 1280x960, 4xAA/16xAF, performance
results were not as severely affected with the K8N Diamond Plus at the higher AA
and AF settings.
Doom3 was used for OpenGL testing. The game runs smooth under both system
setups with the K8N Diamond Plus using maximum in-game settings. Vsync was
disabled and High Quality was selected in the driver panel.