14 DAY TRIAL // We've just come across some benchmark results that really make us hope that that issue with washed-out graphics in Mantle is just a post-processing flaw instead of something inherent in the application programming interface.
For those that aren't up to speed with recent developments in the field of gaming APIs, it has been found that Mantle makes graphics look a bit washed out compared to Direct3D, although performance is definitely better.
It was theorized that the game only needed a patch to overcome that minor issue though, and recent benchmarks should be enough to make everyone wish that to be true.
After all, it would be a shame for a graphics glitch to mar the record of the API when performance is so much better.
The folks at TechReport.com have provided some benchmarks of Battlefield 4 run on two different systems.
First off, there was a benchmark test run on a computer featuring Core i7 4770k coupled with GTX 780 Ti graphics and, later, R9 290X (Mantle On and Off).
Then, the same tests were run on the A10-7850K accelerated processing unit (APU) coupled with the GTX 780 Ti, R9 290X (Mantle On and Off).
Battlefield 4 ran the best on the Core i7 with GTX 780 Ti, unsurprisingly, since those are the fastest beasts on the gaming market at the moment.
Still, Mantle adds 12 frames to the FPS when in a GPU-bound scenario, but in a CPU-bound scenario, FPS goes up by 41 frames. Thus, Mantle allows the R9 290X to only lag 15 frames behind the GTX 780 Ti.
Latency tests paint Mantle in an even more glowing halo, reducing 99th percentile frame time in CPU-bound scenarios, though in GPU-bound scenarios the GTX 780 Ti still wins.
In layman terms, this means that Mantle is really as amazing as AMD claims, especially since the tests would have been more relevant if a Core i5-4430 CPU was used instead of a Core i7, and a GTX 780 was employed instead of the Ti.
