by Alexey Stepin , Yaroslav Lyssenko
05/21/2009 | 08:06 PM
Although we usually blitz-test all new Catalyst driver versions right after their official release, we made an exception for version 9.4 because the developers claimed no performance improvements in it compared with the previous Catalyst version 9.3. So, that driver version was of no particular interest to gamers, although it did fix certain image quality issues in some games. Namely, it eliminated the flickering in World of Warcraft on CrossFireX configurations and artifacts in Age of Conan in DirectX 10 mode.
The recently released Catalyst 9.5 promises much more improvements. According to Advanced Micro Devices, the new driver should ensure a significant performance improvement in a few situations when it gets limited by the system CPU. Among the pointed out gaming titles and tests are the following ones:
Although the first one of these three is a purely synthetic benchmark and is of no significant interest to gamers, we decided to keep it among our test applications. Moreover we also added a few games that haven’t been mentioned in the release notes. The complete list of games and applications used in our today’s test session will follow below.
Besides the promised performance improvements, new Catalyst 9.5 also contains a number of fixes. In particular, it will please those of you who encode video, because version 9.5 is free from the existing ATI Video Converter issues and also supports 64-bit Windows Vista OS. You can now use any Radeon HD 4800, 4700 or 4600 models for work with video converter.
They have also fixed a number of bugs in the driver and catalyst Control Center. Here are a few of the fixed issues for Windows Vista OS:
You can check out the complete list of fixes in the new Catalyst 9.5 driver version on the official Advanced Micro Devices web-site, on the page devoted to ATI Radeon HD graphics cards drivers. This document also contains the list of potential problems that the user of Catalyst 9.5 driver can theoretically face, but they are mostly related to video playback and work with HDTV devices. As for us, we are primarily interested in the new driver performance in games. So, it is time to check out how gamers may benefit from upgrading their driver with the latest Catalyst 9.5.
We are going to check out the performance of the new ATI Catalyst driver version 9.5 using the following testbed:
Both versions of ATI Catalyst driver were configured 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:
We also tested Radeon HD 4890 in CrossFireX configuration. This way we could estimate the improvements brought by the new Catalyst 9.5 driver not only in the common single-GPU mode but also in multi-GPU mode.
We ran our tests in the following resolutions: 1280x1024, 1680x1050, 1920x1200 and 2560x1600. We didn’t test Radeon HD 4770 in the latter one. 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.
Our CPU can hardly be a bottleneck in this case limiting the performance. Nevertheless, the promised 10% performance gain in CrossFireX mode is there, although we could only achieve it in 1280x1024 resolution. In higher resolutions the gain was only 4-6%.
I would also like to say that we also registered a slight performance improvement for the single Radeon HD 4890: in 1280x1024 it sped up by about 7% and in more resource-hungry screen modes – by 4% at the most. The budget Radeon HD 4770, on the contrary, lost the same 4% of speed in the same resolution. Therefore, the owners of these graphics cards should think twice before upgrading their Catalyst driver with version 9.5, since the performance is already quite low and each frame counts.
During the tests in Company of Heroes we faced some problems. By default the game forces “vsync” option that is why maximum performance depends on the screen refresh rate, which is 60fps in our case. Theoretically, it can be disabled using “novsync” argument on RelicCoH.exe launch. However, every time we tried to do it, the game lost stability and crashed when we tried changing the resolution.
For the reasons mentioned above we couldn’t find out if the performance of Radeon HD 4890 CrossFireX tandem did in fact improve with Catalyst 9.5, but single-GPU graphics accelerators did show stable increase from 1% to 5%. Well, it is somewhat lower than the 10% ATI has promised us, however, our CPU is not one of those models that could limit the performance in this game and the 10% gain was promised for these particular processors. I would also like to point out slight increase of the minimal performance in low resolutions demonstrated by Radeon HD 4890. Nothing like that can be observed by Radeon HD 4770.
They didn’t declare any performance changes for Crysis Warhead and our tests proved exactly that: we saw no performance difference between Catalyst 9.4 and Catalyst 9.5 drivers.
We included this game into our today’s test session because its engine uses API OpenGL. The only difference from the performance with Catalyst 9.4 was a slight average performance drop by Radeon HD 4890 CrossFireX tandem in 1280x1024 that didn’t exceed 3% at the most. Quite logical, since ATI didn’t say anything about improving the performance of the driver part responsible for OpenGL support.
We decided to use this game as one of the few titles supporting DirectX 10.1. The release notes for Catalyst 9.5 don’t mention it in any way that is why we didn’t anticipate anything especially interesting here. However, Radeon HD 4890 CrossFireX tandem suddenly demonstrated a pretty serious gain, about 30%, and the subjective level of gaming comfort has also increased significantly. We don’t know what specific optimizations have been put in place there; the improvements may have been applied to the algorithms for work with video memory or CrossFire logic. One thing, however, is certain: the H.A.W.X. fans who own two radeon HD 4890 cards and 30” monitors will be very pleased with this performance gain. We would strongly encourage them to upgrade to Catalyst 9.5.
Unfortunately, this benchmark doesn’t support 2560x1600 resolution, however, it become the only test where all participants performed stably faster in all resolutions. The biggest performance gain belongs to Radeon HD 4890 CrossFireX and reaches 10-14%. Single-GPU solutions performed a little more modest and gained around 7-12%. However, it is more than enough for ATI’s promises to be considered fulfilled.
Trying to minimize the CPU influence in 3DMark Vantage we used the Extreme profile (1920x1200, 4x FSAA and anisotropic filtering).
Looks like ATI does in fact pay most attention to optimizing the functioning of CrossFireX technology. Although this time they didn’t claim any improvements in 3DMark Vantage, Radeon HD 4890 CrossFireX tandem reached a noticeably higher total score with Catalyst 9.5 than with version 9.4.
The results of individual benchmarks show that the gain actually happens only in the first one of them and is most significant in 1920x1200 resolution reaching 9%. In all other cases the performance difference between the two Catalyst driver versions is no more than 1-3%.
Our express-test of ATI Catalyst 9.5 drivers not only confirmed that the performance does in fact improve in the games and benchmarks specified in the driver release notes, but also revealed some unexpected improvements in a few additional applications, such as Tom Clancy’s H.A.W.X. and 3DMark Vantage.
Just like in our previous articles of the same kind, at this point we can conclude that ATI still sticks to its promises: we see performance improving where it is supposed to. Although it is not always up to the specified levels, it is important to remember that we are talking about “up to 10%” gain, and not “10%+”. Moreover, we have to repeat once again that the CPU we use in our testbed is definitely not one of those solutions that may have any more or less serious effect on gaming performance.
As for the general tendencies, the programmers from AMD graphics division who are responsible for Catalyst drivers and Catalyst Control Center development are evidently focusing their main efforts on improving the performance of the CrossFireX technology, although single-GPU solutions also get some. Moreover, they keep fixing a lot of small and not very small issues in the new Catalyst and CCC versions, and even though these issues are not fatal, they are in fact pretty annoying. All of the fixes are described in detail in the driver release notes and it is certainly a great thing.
We didn’t detect any serious performance drops on transition from Catalyst 9.4 to Catalyst 9.5. The only exception is a 4% drop Radeon HD 4770 showed in BattleForge game in 1280x1024 resolution. Therefore, we have every reason to recommend all owners of contemporary ATI Radeon HD 4000 graphics cards to upgrade to the new Catalyst 9.5 driver.