Gigabyte V-Tuner and RADEON 9200: Forget About Overclocking

The CD-R disk we received from the manufacturer contained an utility called V-Tuner. It is a system monitoring and overclocking tool. Since RADEON 9000, unlike its elder, PRO, brother, is locked against overclocking, we supposed that Gigabyte Maya II R9200 will also be impossible to overclock. And we were not mistaken: Gigabyte’s utility was a bright, but useless toy. We also tried to overclock the card with the latest version of RivaTuner, but to no avail. Increased frequencies of the memory and graphics chip didn’t tell on the benchmarks results: the card’s BIOS seemed to be resetting them to the nominal values automatically. Probably, V-Tuner is intended for other cards that don’t have anti-overclocking protection. In fact, this protection system seems proper here as the graphics chip has no active cooler and memory chips are not cooled at all. If there were no protection system like that, an unsophisticated user could overclock the card to death.

As for image quality, it was good enough in all modes supported by our display (we used Hitachi CM772). In any case, we noticed no difference as compared to TYAN Tachyon R9500 PRO.

Overall, Gigabyte Maya II R9200 graphics card leaves a nice impression, although boasts nothing extraordinary except attractive looks.

Testbed and Methods: Value Cards Battle

We decided to test our Gigabyte Maya II R9200 together with RADEON 9000 PRO and RADEON 8500 based cards from ATI and the value graphics solution from NVIDIA: GeForce4 MX440. As overclocking is impossible, we ran all the cards at their regular frequencies.

The following system was used as a testbed:

• AMD Athlon XP 2600+ (Thoroughbred) CPU (2.083GHz, 166MHz (333MHz) FSB);
• EPoX EP-8K3A mainboard;
• 512MB PC3200 Corsair XMS3200 (2-2-5 1T, 166MHz (333MHz DDR));
• Maxtor DiamondMax Plus D740X 40GB HDD;
• Microsoft Windows XP SP1;
• Drivers: VIA Hyperion 4-in-1 v.4.47, ATI Catalyst 3.4 (for ATI cards), NVIDIA Detonator 44.03 (for GeForce4 MX440).

We used the cards in the AGP 4x mode. To prove that AGP 8x brings no performance boost, we tested Gigabyte’s card in Unreal Tournament 2003 and 3DMark 2001SE on an i875P-based platform (Pentium 4 2.40C). For a better comparison, we also included Quake 3 Arena. You can see the results of AGP 8x tests in the end of the review.

Here’s the list of our benchmarks:

• Futuremark 3DMark 2001SE Build 330;
• Unreal Tournament 2003 v2225, Antalus Flyby;
• Quake 3: Arena v1.32, Demo four;
• Return to Castle Wolfenstein v1.4, Demo checkpoint;
• Serious Sam: Second Encounter v1.05, The Grand Catherdral;
• Jedi Knight 2: Jedi Outcast  v1.04, Massasi Temple – Lightsaber Test;
• Splinter Cell v1.2b, 1_1_1Tbilisi Demo.

We excluded Codecult CodeCreatures Benchmark Pro and Futuremark 3DMark03 as they are too difficult for value graphics solutions to pass.

Every gaming benchmark was run with the highest graphics quality settings. We set 1024x768, 1280x1024 and 1600x1200 resolutions in the “raw speed” mode and 800x600, 1024x768 and 1280x1024 resolutions in the 2xAA+8xAF mode. We didn’t enable 4x anti-aliasing, as value cards do not provide enough gaming performance in this mode for comfortable playing, especially in modern games.