Unless you have been living under a rock the last several weeks, you would know that Sweeney and company have finally released Unreal Tournament 2003. I was definitely interested in seeing just how well the Parhelia would deal with a game that's as graphically intense as this one. I think it's safe to say that this will remain true until until Doom III ships.
I've seen a lot of people post on the net that this would be one of those titles that make the Parhelia look pretty bad, so let's see if we can separate fact from fiction.
Unreal Torunament 2002 was tested using the following settings:
Surround Gaming - 2400x600 resolution
Normal - 1024x768 resolution
Trilinear filtering forced via driver
16X FAA enabled where noted
In-game graphics settings at maximum quality
A couple of comparison screenshots to check out.
All benchmarks were taken using the flyby benchmarking feature with sound disabled. I decided not to test any of the botmatch benchmarks as they're more dependent on the CPU.
Unreal Tournament 2003 Performance Average Frame Rate
Surround Gaming with FAA
First, let's talk about the overall numbers. Again, we're talking about a brand new title that hasn't yet been fully optimized and I suspect the same could be said of the drivers as well.
In terms of normal gameplay, the Parhelia has no problems across the board. When analyzing Surround Gaming under the more intense maps such as Antalus we can see that the numbers, in my opinion, are acceptable. When dealing with less intense maps the Parhelia has no problems.
As far as FAA is concerned, you've got to be careful about which maps are being used. The Antalus map seems like a worst case scenario, while some of the others are more mainstream. You can get probably get away with using FAA, but there are a few maps that are just too much for both FAA and Surround Gaming.
The Parhelia is capable of delivering good performance regardless of whether or not you plan to use Surround Gaming or not. Quite frankly, I was surprised by the Surround performance. Why? Because, a lot of the numbers that were tossed around while the game was still under development seemed to paint a less than favorable picture of the Parhelia. This is precisely why I believe that no unreleased software should ever be used to evaluate a product. Not only doesn't the average Joe not realize what is being rendered, but it also doesn't accurately portray final performance levels.