The test was conducted using AMD’s Catalyst 14.2 Beta driver (14.3 was just released a few days ago).
So how did the Mantle API fare against DirectX 11?
CPU : AMD A10-7850K (Kaveri) APU
Motherboard : MSI A88X-G45 GAMING
Graphic Card: Sapphire Radeon R9 280X Vapor-X 3GB OC
Memory : Corsair Vengeance 1600Mhz 8GB
Storage : WD Blue 320GB
Power Supply : FSP Raider 650w
Cooling System : Corsair H55 Water Cooling Kit
Casing : Corsair Graphite 230T
OS : Windows 8 Pro
The test uses Battefield 4’s built-in Frame Time log recording, executed using “PerfOverlay.FrameFileLogEnable 1” in the console. I used Baku mission, recording it right as the start (“Total Eclipse of the Heart”) up till the part where the crew is about to pass through the tunnel.
1920 x 1080 resolution, details set to Ultra.
🙁 Baku the only scene I have even though I’ve nearly completed the campaign (gaaahhhh I have to replay it!)
The log was then analyzed using the BF4 – Frame Time Analyzer
This is the result for Mantle API.
And this is the result for DirectX 11.
The AMD A10-7850K APU is a processor with pretty good performance already, and with the R9 280X the performance gap isn’t great but nevertheless it’s a start.
What I’m concerned is that while the Mantle API gave a frame rate boost, it also has the lower minimum frame.
I’ll probably come up with a Revision 2 of the Mantle vs DirectX test. Perhaps the next one will be with DirectX 12, and the setup would most likely be a lower range processor (eg A10 7700K or even A4 4000) with at least an R9 290 or R9 280X Crossfire.
If you want to try out the above setup with Battlefield 4 – head over to Low Yat Plaza next week, goldfries.com’s system will be part of AMD’s showcase units. 😀