It's been a while since we last benched AMD Catalyst drivers which is our fault and not AMD's. Anyway AMD repleased the Catalyst 12.9 Beta two weeks ago on the 26th of September so we are going to test it and compare it with no less that 12 previous Catalyst drivers. Spoiler Alert: As soon as we started to test the 12.9 Beta we noticed a performance increase over the previous 12.8 WHQL which is always a good sign.
Article in English |
Artikel in Deutsch |
Article en français |
Articolo in italiano |
Introduction
Find all the benchmarks/games results on the following pages.
Discuss this article in the forums [pagebreak]
Test Setup
Hardware
| Mainboard |
|
| CPUs |
|
| Memory |
|
| Graphic Cards (Driver) |
|
| Drivers |
|
| HDD |
|
| PSU |
|
Discuss this article in the forums [pagebreak]
Benchmarks and Games Settings
Futuremark
- 3DMark 11
(1.0.3)
- GPU Test - Performance Preset
- 3DMark Vantage (1.1.0)
- GPU Test - Performance Preset
Unigine Heaven DX11 Benchmark
| Version | 3.0 |
| API | DirectX 11 |
| Tessellation | Extreme |
| Shaders | High |
| Anisotropy | 16x |
| Stereo 3D | Disabled |
| Anti-aliasing | 8x |
| Full Screen | Yes |
| Resolution | 1920 x 1080 |
Geeks3D TessMark
| Version | 0.3.0 |
| Map Set | Set 4 (2048 x 204 |
| Tesselation | Insane (x64) |
| Full Screen | Yes |
| Resolution | 1920 x 1080 |
| Anti-aliasing | 8x |
| Tessellation | High |
| Duration | 60000 (ms) |
Geeks3D FurMark (Burn-in test)
| Version | 1.10.0 |
| Full Screen | Yes |
| Resolution | 1920 x 1080 |
| Anti-aliasing | 8x MSAA |
Alien vs Predator DX11 Benchmark
| Texture Quality | Very High |
| Shadow Quality | High |
| Hardware Tesselation | Yes |
| Advance Shadow Sampling | Yes |
| Full Srceen Anti-Aliasing Samples | 4x |
| Anisotropic Filtering | 16x |
| SSAO | On |
| Screen Resolution | 1920 x 1080 |
Batman: Arkham City Integrated Benchmark
| FullScreen | Yes |
| Resolution | 1920 x 1080 |
| V Sync | No |
| Anti-Aliasing | 8x MSAA |
| Stereoscopic 3D Status | No |
| DirectX 11 Features | MVSS And HBAO |
| DirectX 11 Tesselation | High |
| Detail Level | Extreme |
| Dynamic Shadow | Yes |
| Motion Blur | Yes |
| Distortion | Yes |
| Lens Flares | Yes |
| Light Shafts | Yes |
| Reflections | Yes |
| Ambient Occlusion | Yes |
| Hardware Accelerated PhysX | Off |
Call Of Duty Modern Warfare 3 (5 Minutes GamePlay)
| Resolution | 1920 x 1080 |
| FullScreen | On |
| Vertical Sync | Off |
| Field of view | 70 |
| Motion Blur Amount | Max |
| Graphics Quality | Ultra |
BattleField 3 (5 Minutes GamePlay)
| Resolution | 1920 x 1080 |
| Aspect Ratio | Wide 16:9 |
| Anti-Aliasing | 4x |
| Screen Refresh Rate | 60 Hz |
| Sync Every Frame | No |
| Shadows | Yes |
| Specular Map | Yes |
| Depth Of Field | Yes |
| Screen Space Ambient Occlusion | High Quality |
| Soften Smoke Edges | Yes |
| Bullet Impacts | Yes |
| Image Quality | Ultra |
Crysis 2 DX11 + HiRes Texture Pack Benchmark
| Quality | Ultra |
| Resolution | 1920 x 1080 |
| Anti-Aliasing | 4x |
| Map | Central Park/Adrenaline |
| Runs | 1 |
| API | DirectX 11 |
| Edge AA | Edge AA |
| Hi-Res Textures | On |
DIRT 2 Integrated Benchmark
| Resolution | 1920 x 1080 |
| Refresh Rate | 60 |
| MultiSampling | 8x MSAA |
| VSync | Off |
| Aspect Ration | 16:9 |
| Details | Ultra Preset |
Dragon Age II (5 Minutes GamePlay)
| Resolution | 1920 x 1080 |
| Full Screen Mode | Yes |
| Renderer | DirectX 11 |
| Vertical Sync | no |
| Graphics Detail | Very High |
| Anti-Aliasing | 8x |
| Anisotropic Filtering | 16x |
| Screen Space Ambient Occlusion | Yes |
| Diffusion Depth of Field | Yes |
| High Quality Blur | Yes |
The Elder Scrolls V: Skyrim (5 Minutes GamePlay)
| Resolution | 1920 x 1080 |
| Aspect Ratio | Wide 16:9 |
| Anti-Aliasing | 8x |
| Anisotropic-Filtering | 16x |
| Details | Ultra |
| Vertical Sync | Disabled |
S.T.A.L.K.E.R. Call of Pripyat Benchmark
| Preset | Ultra |
| Renderer | Enhan. full dyn. lighting (DX11) |
| Video Mode | 1920 x 1080 |
| MSAA | 4x |
| MSAA for A-tested objects | DX10.0 Style [Standard] |
| SSAO Mode | HDAO |
| SSAO Quality | High |
| Use DX10.1 | No |
| Enable Tesselation | Yes |
| Contact Hardening Shadows | Yes |
StarCraft II (5 Minutes GamePlay)
| Display Mode | Fullscreen |
| Refresh Rate | Default |
| Resolution | 1920 x 1080 |
| Vertical Sync | No |
| Texture Quality | Ultra |
| Graphics Quality | Ultra |
Discuss this article in the forums [pagebreak]
3DMark 11
3DMark Vantage
Discuss this article in the forums [pagebreak]
Unigine Heaven
TessMark
Discuss this article in the forums [pagebreak]
Alien vs Predator
Batman: Arkham City
DIRT 2
Discuss this article in the forums [pagebreak]
Call of Duty Modern Warfare 3
BattleField 3
Discuss this article in the forums [pagebreak]
Crysis 2
Dragon Age 2
Discuss this article in the forums [pagebreak]
S.T.A.L.K.E.R. Call of Pripyat
The Elder Scrolls V: Skyrim
StarCarft 2
Discuss this article in the forums [pagebreak]
Power Consumption
Discuss this article in the forums [pagebreak]
Benchmark Summary
Synthetic Benchmarks
Games
Discuss this article in the forums [pagebreak]
Performance Index
Discuss this article in the forums [pagebreak]
Conclusion
When
looking at the performance offered by the Catalyst 12.9 Beta one can see that it is much like the performance of the Catalyst 12.7 Beta
from june 25. Last time we were a bit disappointed by the Catalyst 12.8 WHQL
drivers because of the three following things:
- Performance under Dragon Age II went from 50 fps average to 37 fps
- Batman Arkham City was freezing loading
- Power consumption under load was 20 watts higher than with previous drivers.
AMD seems to have fixed those drivers issues as this time we didn't notice
anything abnormal at all. They even are the best drivers we ever tested on the
Radeon HD 7970 on average. Compared to the 12.8 WHQL we saw an improvement lower
than 1 % everywhere but in 3DMark Vantage where it's the first time we score
that high using stock settings, 32'040 points. AMD fixed (again) the performance
problem under Dragon Age II we saw the first time using the Catalyst 12.xB from
June 25, and again with the Catalyst 12.8 WHQL. The power consumption too is
back to normal with 250 Watts under heavy load against the 270 Watts we measured
last time (12.8 WHQL).
Did AMD listened to us? Not that they only fixed the issues we spotted
last time but it looks like they also made sure that the 12.9 Beta will never
perform lower than the 12.8 WHQL for our test setup as we didn't measure any
decrease in performance.
Author: Christian Ney c.ney@ocaholic.ch
Discuss this article in the forums