Benchmarks II
While it was difficult to have anything to compare the Scan system with on 2D benchmarks, 3D benchmarking opens up single system comparisons. The idea is to run our series of gaming tests at 1024x768x32. However, we'll also run the same tests with 4x antialiasing and 8x anisotropic filtering employed. That'll inform us as to whether the Gainward FX5200 Ultra has what it takes to run smoothly with image enhancement settings. The very fact that the Sony 15" TFT cannot run at more than 1024x768 resolution makes the comparison even more interesting. 3DMark 2001SE kicks off proceedings.
The standard benchmark score of 10,387 is heavily influenced by the subsystem. A couple of tests, Lobby and Car, require masses of geometry to function well. Tests such as Dragothic and especially Nature put massive load on the video card. The FX 5200 Ultra is a midrange card at best. That's why it suffers heavily when bandwidth-sapping AA and AF are used.
Serious Sam: TSE now. The lovely Sierra De Chiapas timedemo. Extreme Quality setting was used.
Fluid and acceptable with standard settings, but it soon grinds to a crawl in places with a heavy load. It's a little unfair to tax the card so heavily, because there are a number of intermediate image enhancement settings. Remember that sound is always disabled for our gaming benchmarks.
Comanche 4 now.
The performance drop is limited to less than 50% for the first time - just. The basic speed is quite good, though.
Unreal Tournament 2003 demo's Flyby now.
It's a little difficult to believe that the card can lose so little in this antialiasing and anisotropic filtering nightmare of a test. The 44.03 drivers may be playing optimising tricks here.
Quake III wraps it up.
Another bashing with AA and AF applied. The power of the card still allows for decent, fluid play with the best settings used. Our bottom line, with respect to the system's 3D ability, is to stick to either no AA and AF or 2x AA / 4x AF at best. A faster card would change our observations, naturally.
