ATI Catalyst 9.3 Driver: Performance Express-Test

March is here and the graphics division of Advanced Micro Devices released a new version of ATI Catalyst drivers. As usual, we performed a quick blitz-test aimed at revealing the performance improvement delivered by the new driver version.

by Alexey Stepin , Yaroslav Lyssenko
03/18/2009 | 11:28 PM

Introduction

ATI Catalyst drivers development follows a very precise schedule. The number after the point in the driver index stands for the release month. Since February is over and March has already passed its middle, the time has come for the new ATI Catalyst driver version 9.3 to be released. So, what do the ATI Catalyst driver developers can tell us about the new version?

 

According to the release notes, Catalyst 9.3 doesn’t promise any performance breakthroughs in any of the contemporary games, except Lost Planet: Colonies shooter, where we should see up to 20% performance improvement for Radeon HD 4800 cards and up to 50% improvement for less powerful solutions, such as Radeon HD 4600, 4500 and 4300. They also mentioned that performance may increase in a few other cases when it is limited by the CPU. They also haven’t forgotten about Folding@Home distributed computing project fans: the performance in GPGPU mode is also supposed to be increased.

As usual, the new Catalyst release corrects a number of bugs and issues. This time, they are mostly connected with video playback in Windows Vista:

Besides, Catalyst 9.3 is the first unified ATI driver supporting Windows Vista and Windows 7 operating systems, although the latter is still under development and will see the light of day no sooner than in 6 months. This way AMD is trying to say that their graphics division is the leader in developing drivers for the upcoming OS that will replace Vista. According to the company officials, Catalyst 9.3 provides complete support for WDDM 1.1 in Windows 7. Moreover, the games should work faster under the new OS than under Windows Vista. We are not going to check out if these statements are true and correct within our today’s article, but we will return to this topic a little later and offer you a massive comparison of gaming performance under these two operating systems side by side.

We also have to say that Catalyst 9.3 is the last ATI driver to support ATI old Radeon lineups. Starting with the next version, the updates for Radeon X1x00 will be made on a quarterly basis, just like it used to be done before. This way the company is going to cease supporting the inherited solutions and will invest all their efforts into maximum support for the contemporary ATI Radeon models. Well, it is time to check out how gamers can benefit from installing the new Catalyst 9.3.

Testbed and Methods

We are going to check out the performance of the new ATI Catalyst driver version 9.3 using the following testbed:

Both versions of ATI Catalyst driver were configured to provide the highest possible quality of texture filtering with minimal effect from default software optimizations. We enabled transparent texture filtering. As a result, our driver settings looked as follows:

This time we used the following games for our performance tests:

We selected the highest possible level of detail in each game using standard tools provided by the game itself from the gaming menu. We didn’t use any non-standard settings, except a few situations, described separately later in the article.

For our tests we chose two most significant representatives of ATI Radeon HD 4800 family:

This choice was determined by our intention to study the peculiarities of the new driver version with classical single-GPU architecture, as well as with a dual-processor solution based on CrossFireX technology.

We ran our tests in the following resolutions: 1280x1024, 1680x1050, 1920x1200 and 2560x1600. We added MSAA 4x antialiasing to the standard anisotropic filtering 16x in all tests. We enabled antialiasing from the game’s menu. If this was not possible, we forced it using the appropriate ATI Catalyst driver settings.

Performance was measured with the games’ own tools using the original demos. We measured not only the average speed, but also the minimum speed of the cards if possible. If there were no appropriate tools available in the game itself, we used Fraps utility version 2.9.8 to record our performance measurements in this game. In the latter case we ran the test three times and took the average of the three readings for the performance charts.

Performance

Enemy Territory: Quake Wars

We didn’t detect any significant performance changes compared with Catalyst 9.2. However, no one promised that Catalyst 9.3 will deliver any improvements in OpenGL, so the obtained results are quite expected. We did notice a slight performance drop by Radeon HD 4870 X2 in 1280x1024. However, it only made 1%, which lies within the allowed measuring error.

Fallout 3

Fallout 3 is not mentioned in the release notes for Catalyst 9.3 and our tests showed that there is no difference in performance between Catalyst 9.3 and Catalyst 9.2 in this game. Despite the fact that we used Fraps in manual mode, the results coincided with up 1.2 frames per second precision.

Crysis Warhead

The average performance remained the same even with the new ATI driver, but we did detect slight increase in the minimal speed in 1280x1024 resolution. However, taking into account that we used Fraps in the manual mode for testing, it could be some accidental fluctuation, because we can’t physically repeat the test run exactly in this case and multiple factors may affect the final result, such as random HDD requests, for instance.

F.E.A.R. 2: Project Origin

We didn’t detect any differences between Catalyst 9.2 and 9.3 versions in this game besides a slight increase in the minimal performance.

Lost Planet: Colonies

We did see the performance improvement in the only game where they actually promised it. It didn’t reach the anticipated 20%, however, in 1280x1024 it hit a pretty good mark of 11% for the dual-GPU card and 8% for the single-GPU one. In 1680x1050 the boost lowered to 8.3% and 7.7% respectively. In 1920x1200 Radeon HD 4870 X2 sped up by the good 14%, while Radeon HD 4870 – only by 2.2%. And in the highest 2560x1600 resolution only the X2 model benefited from the new driver version, while its single-processor counterpart didn’t demonstrate any serious performance increase.

Not bad, although it is evident that ATI driver developers focus mostly on improving the performance of the multi-GPU solutions. Nevertheless, the owners of regular single-GPU ATI Radeon HD solutions will not be forgotten either. I would also like to say that systems equipped with less powerful CPUs may demonstrate an even greater performance improvement reaching the 20-50% ATI have promised us.

World in Conflict

World in Conflict also wasn’t mentioned in the Catalyst 9.3 driver release notes, so there is no evident performance increase there. Just like in all the previous cases, we didn’t detect any performance drops or image quality worsening.

Futuremark 3DMark Vantage

As usual, trying to minimize the CPU influence in 3DMark Vantage we used the Extreme profile (1920x1200, 4x FSAA and anisotropic filtering).

Installing new ATI Catalyst driver didn’t have any effect in any of the 3DMark Vantage tests. The differences between the results obtained with two versions of ATI Catalyst driver are minimal and lie within the benchmark measuring error.

Conclusion

Our express test session with the new ATI Catalyst 9.3 driver didn’t reveal any serious effect on performance in contemporary games, except for Lost Planet: Colonies. It is the only game where a significant performance improvement was promised right from the beginning in the driver release notes. But, I guess, the summarizing charts will speak for themselves:

We didn’t detect any noticeable performance improvement that would evidently exceed the measuring error acceptable for our testing methodology in any of the games except Lost Planet: Colonies. Nevertheless, we have to admit that once again ATI kept their promise: we did see performance growth where they promised we would. Just like during our previous test session last month the biggest performance improvement occurred for dual-processor ATI Radeon HD solutions, while classical single-GPU products benefited much less, especially in high and ultra-high resolutions. In other words, ATI driver developers are evidently focusing all their effort on making the best out of CrossFireX technology. In fact, it is not surprising at all, especially if we recall that the graphics division of Advanced Micro Devices bet big on this particular technology in the fight for title of the highest-performance 3D graphics solutions maker. Unfortunately, we couldn’t confirm or deny ATI’s claims about improved performance in Folding@Home project because TOC F@H Bench that we use wouldn’t work with the new Catalyst driver version. But we are going to get back to this topic later on, so stay tuned.

Summing up, we can state that there is nothing to discourage you from installing the new ATI Catalyst 9.3 driver at this time. It may not deliver too many benefits to the owners of regular gaming platforms based on Windows Vista OS, but it is definitely a more progressive version and the first ATI’s unified driver that supports both: the current Windows Vista and the upcoming Windows 7 operating systems. Besides, as usual, the new version fixes a number of serious issues, such as the ones connected with video playback, for instance.