By volt - April 12, 2003

Unreal Tournament 2003: Performance

Epic has been very successful releasing original Unreal Tournament. Although some people prefer Quake III over UT, the game has gained a lot of respect and fans all over the world. It's been played over and over again at LAN parties. Over the past years the game engine became obsolete (I can't say that about the gameplay though). Gamers wanted something new and visually stunning.

The hype for UT2003 grew strong. The demo has been delayed few times, but we all tend to forget about those things especially that the game has been out for quite some time. So what graphical innovations does UT2003 bring in to the table? First and foremost, a redesigned game engine which utilizes todays graphics cards, including our Gainward 650 XP. Although it's a DX7 based game, a DX8.x based card is recommended.

We had benchmarked two maps at two modes: 1024x768 and 1280x1024.

For reference, our test bed:

P4: 1.6A @ 2.39GHz / 512MB PC2100 / Stock Gainward Ti4600
AMD: Athlon XP 1900+ @ 1.6GHz / 512MB PC2100 / Stock Gainward Ti4200

DM-Asbestos:

DM-Phobos2:

Okay, maybe we wen't a bit crazy with the benchmarks here, but hey! Mode comparisons are always good (especially if you can't find the one that suits your needs). Since this is a widely played game we wanted to test as many modes as possible. Now the gameplay. A bit different than original UT, more like Quake III but nevertheless ass whoopin'. You now have around 50 characters to choose from, more weapons, awsome looking levels and new game types. Visually, the game looks amazing. Every object is crafted to its perfection with maximum number of polygons (without sacrificing the performance).

There is no doubt that this game runs like a charm on our lab card. Even though you see lower numbers, don't be turned by them. It's a different story when you actually play it. The game is CPU limited, that's a fact. There are plenty of settings you can lower / disable (without degrading the quality) in order to get great frame rates.

Without applying any IQ modes, in our first set of benchmarks we get around 140 frames per second (1024x768) and 96 at 1280x1024. Performance changes when switching to 2X Antialiasing. Is it noticeable ? Not really since we still get around 100 FPS on both maps. This is not exactly true for higher resolutions and Antialiasing modes. But we will take care of this in one minute with a little explanation. Remember...fillrate is your friend! Now I don't really recommend playing at high resolution with all the eye candy enabled. As you can see with other tests the performance isn't very good (At least not with this mid-range graphics card). The ultimate setting for UT2003 would be a resolution of 1024x768 with Quincunx AA and 2X Anisotropic filtering enabled You should be able to get an average of 45 on all maps.

Remember our friend fillrate? The ultimate bottleneck for higher resolutions and Antialiasing/Anisotropic filtering.

Fill rate is the rate at which pixels are drawn into the screen memory. Fill rate is a common measure used to illustrate the pixel processing capabilities of todayís 3D graphics processors. Fill rate is usually measured in millions of pixels/sec. (Mpixels/sec.) In 1997, 50-70 Mpixels/sec. was considered state of the art. In 2002, the leading 3D graphics processors will be capable of more than 1200 Mpixels/sec. While this improvement is an incredible achievement, it is still barely enough to create a compelling 3D environment. Rendering pixels at such a high rate consumes enormous amounts of memory bandwidth.

Depth complexity is a measure of the complexity of a scene. It refers to the number of times any given pixel must be rendered before the frame is done. For example, a rendered image of a wall has a depth complexity of one. An image of a person standing in front of a wall has a depth complexity of two. An image of a dog behind the person but in front of the wall has a depth complexity of three, and so on. As depth complexity increases, more rendering horsepower and bandwidth is needed to render each pixel or scene. The average depth complexity of todayís graphics applications is two to three, meaning that for every pixel you end up seeing, it gets rendered two or three times by the graphics processor.

Frames per second (fps), or frame rate, refers to how many times per second the scene is updated by the graphics processor. Higher frame rates yield smoother, more realistic animation. It is generally accepted that 30fps provides an acceptable level of animation, but increasing the performance to 60fps results in significantly improved interaction and realism. Beyond 75fps it is difficult to detect any performance improvement. Displaying images faster than the refresh rate of the monitor results in wasted graphics computing power, because the monitor is unable to update its phosphors (or display) that fast, wasting frame rate beyond its refresh rate.

Let us use depth_complexity of 3 as an average...

For 1024x768 resolution:
fillrate = resolution * depth_complexity * frame_rate
fillrate = 1024 * 768 * 3 * 139
fillrate = 327.9 Mpixel/sec

For 1280x1024 resolution:
fillrate = resolution * depth_complexity * frame_rate
fillrate = 1280 * 1024 * 3 * 96
fillrate = 377.5 Mpixel/sec

(that's without any AA and/or AF)

So, the more fillrate your card can push, the better performance you will get at higher resolutions and/or Antialiasing/Anisotropic filtering modes. Depending on the fillrate of your card, you will be able to run at a higher resolution with little performance hit. You can always try reducing your color depth to 16 bit if you like playing at high resolutions. *If* you find that this doesn't do the trick, try disabling trilinear filtering and other texture details.