3D Mark Getting a DirectX 12 Vs Mantle Feature Test ‘Farandole’ – 7.5 Times the Draw Calls of DX11

Author Photo
Nov 24, 2014
Share Tweet Submit

3D Mark is arguably the single most authoritative benchmark in the PC Gaming world. It is somewhat of a tradition for overclockers and builders to run any variant of the benchmark on a new configuration. Therefore something quite interesting was present in the slides from AMD’s ‘Future of Compute’ Event: a feature test to benchmark Mantle API vs DX12 performance: Farandole.

Futuremark 3D Mark Dandia and BalboaA slide from Future of Compute Event – Public Domain @AMD

Related Mafia III Gets Demo, New Story DLC and 50% Discount All at Once

Farandole is Futuremark 3DMark’s Low Level API feature test with 7.5 times the draw calls

They have also revealed the name of the first DX12 feature test (Farandole) which is enroute for 2015-2016 and will be named “Dandia”. There is also going to be a lightning test named Balboa, which I assume is a voxel based feature test for global illumination. The slides in question have been attached. You know what this means though, once the test has been run on every AMD card, you will get an obvious answer on which API handles better.  Since this is a synthetic benchmark, performance of both APIs will be heavily dependent on the optimization the devs do. since this is something which we can never assume to be perfect and something which is usually not equal in  nature, therein lies an inherent problem: Whatever the answer maybe (the winning API), I doubt all gamers will be happy with it. Futuremark is basically just setting itself up for flames with this Farandole test, but something that is nonetheless essential for reference purposes.3DMark Farandole

This also means that DX12 and Mantle API are both able to handle at the very least 7.5 times the amount of draw call load than DirectX 11 which any tech savvy person would be able to see is a mindbogglingly huge jump to come just from a software update. Good on AMD to set the ball rolling on this one. I also want to see how the DX12 Vs Mantle API performance varies under different configurations, clocks and CPU load and this benchmark should provide the ideal reference point. As we revealed a long time ago, Mantel API is not a redundancy anymore. Its a safety net for developers; even those wanting to work primarily in DX12. Porting time and difficulty is minimum in most cases, so it would make sense for most devs to support both platforms. Besides, the fact remains that Mantle API is out right now, while DX12 still has some time to go as Microsoft brushes up its new OS for release. Either ways, its good to see the Industry moving forward.

Share Tweet Submit