AMD Catalyst 12.6 BETA Benchmark

With the release of AMD Catalyst 12.6 Beta, AMD made an official announcement regarding their new schedule for the future Catalyst drivers:

With the release of the AMD Catalyst™ 12.6 Beta driver (“Catalyst”), AMD would also like to announce that we are moving away from our Monthly Catalyst release plan. Our goal is to ensure that every Catalyst release delivers a substantial benefit to our end users; as we have today with the release of the Catalyst 12.6a Beta. We will still continue with the Catalyst naming convention; Catalyst: Year.Month. You just won’t see a new driver every single month. We are confident that this will only benefit the end user; you’ll only need to upgrade to a new Catalyst driver, when it makes sense.

We would also like to announce the introduction of our new AMD Issue Reporting Form. We made a number of improvements to the older Catalyst Crew Feedback form, making it more comprehensive, ensuring high quality feedback. Feedback is very important in every product company’s life, especially in our fast paced environment. This is not just a check-box for us – we take these reports very seriously. We will review every report posted here and investigate every issue encountered. Please use this form whenever you encounter something worth mentioning.

This announcement is sketchy at best. Why? Well, the latest Catalyst drivers came increasingly later; for instance, Catalyst 12.2 was released ~2 weeks later than when it was supposed to. And taking this announcement into account, isn’t AMD able to cope with a monthly WHQL schedule anymore? I wanted more information regarding this change.

They said something like this: “having two different driver development paths is not necessary”; basically they want to remove the monthly WHQL drivers and focus on just hotfix drivers for specific games/apps. I’m afraid this change will only make driver releases even more scarce, unless AMD skips WHQL certification for their drivers and mark each release as a BETA. Without skipping Microsoft’s certification process I don’t see how they will deliver “urgent hotfixes” on time.

Further more, skipping WHQL certification might turn out to be a double-edged sword, while the drivers will be “served” faster, their quality might degrade over time (from various reasons).

Removing monthly WHQL drivers and focusing on hotfix drivers, new features will take much longer to get implemented. I’ve been waiting for DX10+ SSAA for non HD7k cards for almost 4 month now and no hope of that feature ever being implemented.

The way I see it, AMD GPUs owners fall under two categories: 1. the ones with issues with every other game (CFX especially) that want their hotfixes ASAP; 2. the ones with no issues that just want new features. Even with their “two path development process” AMD couldn’t (fully) satisfy either one but with this new release plan the first category of people will get what they want, leaving the second category to slowly boil.

But enough with all my non-sense, let’s see how AMD Catalyst 12.6 Beta performs compared to the latest (last?) WHQL driver, 12.4 WHQL and the best performing driver so far, 8.97 18 April. Apparently, AMD Catalyst 12.6 Beta is not the “politically correct” name, AMD Catalyst 12.6a is…

Another change comes was done in the versioning the 3D Driver.

Prior to installing a driver, a sweep was done in safe-mode using Phyxion Driver Sweeper.

All 3 drivers are tested in 5 synthetic benchmarks and 5 games.

• 1. 3DMark06 1.02 Professional Edition - 1680×1050 – 2xAA – Anti Aliasing Quality = 0 – 4xAF – Shader Model 2.0 = Return to Proxycon + Firefly Forest – HDR/Shader Model 3.0 = Canyon Flight + Deep Freeze – 2 Runs.
• 2. 3DMark11 1.0.3 Professional Edition - 1680×1050 – 2xMSAA – 4xAF – Graphic Tests 1-4 – 2 Runs.
• 3. Unigine – Sanctuary 2.3 – 1680×1050 – 2xAA – 4xAF – High Shaders – Ambient Occlusion = OFF – 2 Runs.
• 4. Unigine – Tropics 1.3 – 1680×1050 – 2xAA – 4xAF – High Shaders – Ambient Occlusion = ON – Reflections = ON – 2 Runs.
• 5. Cinebench x64 11.5 – 3 Runs.
• 6. Crysis 2 1.9 using Adrenaline Benchmark Tool – Extreme Preset – 1680×1050 – 4xAA – Hi-Res Textures – Central Park – 3 Runs.
• 7. Metro 2033 using Update 1.2 Benchmark Tool – Very High Settings – 1680×1050 – 16xAF – No AA for DirectX 9 and 4xMSAA for DirectX 10 and DirectX 11 – 3 Runs.
• 8. DiRT 3 – Ultra Preset – 1680×1050 – 8xMSAA – 3 Runs
• 9. Battlefield 3 – Ultra Preset – 1680×1050 – 4xMSAA – 16xAF – Multiplayer @ Strike at Karkand (32p map, Large Conquest), main street: US Deployment -> B Flag – 3 Runs x 60s.
• 10. Elder Scrolls V: Skyrim v1.5 – Ultra Preset – 1680×1050 – 8xAA – 16xAF – Tested in the area East of Rorikstead – 3 Runs x 60s.

All synthetic benchmarks were ran 2 times except Cinebench (3 runs), all games were ran 3 times.

Table of Contents:

Page 1 – Introduction
Page 2 – 3DMark06,3DMark11,Cinebench and Unigine
Page 3 – Crysis 2, Metro2033, Battlefield 3, DiRT 3 and Skyrim
Page 4 – Conclusion

Page 2 – 3DMark06,3DMark11,Cinebench and Unigine