2D Image Quality, Noise and Overclocking

I cannot find any issues worth commenting in terms of the 2D picture quality produced by both graphics cards: it was simply impeccable in both cases. They drew an acceptable onscreen image in all resolutions up to and including 1600x1200@85Hz. However, this largely depends on the display as well, and you may get different results with a different display.

As for overclocking, Xtasy 9800 PRO didn’t hit the mark set by Club 3D RADEON 9800 PRO and ATI RADEON 9800 PRO cards and worked stably at 445Mhz core and  756MHz (378MHz DDR) memory frequency. Leadtek A350 WinFast A350 TDH VIVO did brilliantly well: 500Mhz and 950MHz for the graphics chip and memory respectively.

Now, let’s say a few words about the efficiency of the cooling systems and the noise they generate during work. We can criticize Xtasy 9800 PRO for absolutely the same issues as the original RADEON 9800 PRO in terms of the produced noise. However, it is not too loud and the cooler does its job quite well.

As for Leadtek WinFast A350 TDH, its massive cooling solution proved to be more efficient: it kept the GPU and memory chips cool even during overclocking. The aluminum case, connected to the heatsink through a thermal gasket must have contributed to it quite significantly. The two fans blowing the air through the graphics card’s case were pretty loud, but not unbearably so. So, we could put up with the noise for the sake of excellent overclocking results.

Testbed and Methods

So, now it’s high time we tried to figure out  how greatly the performance of contemporary High-End graphics accelerators drops as soon as we enable various modes of full-screen anti-aliasing. To make the experiment more illustrative and unbiased, we turned on maximum graphics quality settings in the benchmarks, but disabled anisotropic filtering. We also used the latest versions of the drivers (ATI Catalyst 3.6 and NVIDIA Detonator FX 45.23) available at the time when the test session was actually in progress.

Our testbed configuration looks as follows:

• AMD Athlon XP 2600+ Thoroughbred CPU (2.083GHz, 333MHz FSB);
• EPoX EP-8K3A+ mainboard;
• 512MB Corsair XMS3200 (2-2-5 1T, 333MHz) DDR SDRAM;
• Maxtor DiamondMax Plus D740X (2MB buffer) HDD, 40GB;
• Creative SoundBlaster Live! 1024 audio card;
• Microsoft Windows XP SP1;
• Drivers: VIA Hyperion 4-in-1 v.4.49, ATI Catalyst 3.6, NVIDIA Detonator FX 45.23.

We used the following benchmarking software;

• Quake 3: Arena v1.32
• Return to Castle Wolfenstein v1.4
• Unreal Tournament 2003 v2225 (flyby inferno, flyby antalus and flyby asbestos)
• Star Trek: Elite Force 2
• Serious Sam: The Second Encounter (Grand Cathedral demo and Elephant Atrium demo)
• CodeCult Codecreatures Benchmark Pro

We also wanted to include one more gaming benchmark, Tomb Raider: Angel of Darkness as it is a new game that actively uses pixel shaders version 2.0. However, the developers were late in releasing patch ver.49 for it. The patch (version 42) we had at our disposal at that time provided simply inadequate results.

So, let’s run the tests!