Nvidia overhauled their drivers awhile back as an answer to Mantle and we're seeing the benefits in the benchmarks.
It could allow better use of GPU compute as well, because as I understand it it's a better fit for GPU nowadays.
I don't know if DX12 works on Windows phone but surely that would actually give credence to DX12 as a multi platform solution. OpenGL will remain the API of choice for Linux and Mac but as far as PC gaming is concerned those platforms are not as popular as Windows.I'd have to agree with some of the other posts. DX12 is hardly "multiplatform" in a conventional sense since it is Microsoft only. It's either DX12 on Xbox running Windows 10, or DX12 on Windows 10 PC or DX12 on Window 10 phones.
Meanwhile something like OpenGL family would be a better example of "multiplatform", functioning across mobile, web and desktop platforms, despite its lack of use on consoles and decline for majority of PC gaming. If the combination of complete merging and rewrite of OpenGL & OpenGL ES in glNext is as pictured, your idea of DX12 being "multiplatform" is almost comical.
https://www.khronos.org/assets/uplo...4-siggraph-bof/OpenGL-Ecosystem-BOF_Aug14.pdf
Definitely, it will lower the price of entry for good PC performance. It's going to be even more interesting to compare PC and consoles now, see where do they stack up against various CPUs, GPUs running DX12.
Nvidia and MS wanted that back in 2010.
http://blogs.nvidia.com/blog/2014/03/20/directx-12/
So you're saying that the drivers were the main reason in the performance gap between AMD and NVIDIA. That bodes well for AMD since a 290X is pretty much on par with a 970GTX in terms of performance sometimes even eclipsing it in many games that I've seen benchmarked. It's cheaper and benches better at higher resolutions too, I'm thinking they have the full 4GB's too. Win win for AMD it seems.
So you're saying that ARMA is not possible on consoles? What exactly about ARMA is not possible? What is so revolutionary about it?
Right. Its a huge deal for PC's as it essentially gives PC's the same sort of low level efficiency advantage that consoles have always had. For the XB1, it will be minor API improvements at best.
Basically, yea. If a game is developed with DX12, it should enable them to get fuller use of your CPU, hopefully increasing your CPU performance ceiling.
In CPU bound scenarios. Hence the gap between AMD DX11 and Mantle.
I think it's a solid card and a valid alternative to the 970. But power consumption is up.
I think it's a solid card and a valid alternative to the 970. But power consumption is up.
At least AMD and DICE accomplished part of what they set out to do, draw attention to CPU utilization and low-level access. Before Mantle IIRC, nobody was talking about those seriously.
The motivation for AMD is to entice devs/studios to use Mantle to cross port and still reap the benefits of low(er) layers. The whole idea behind is to leverage their position in the console landscape.
And they would be foolish not to try that. It makes a great deal of sense. Technically Mantle could be considered a "standard" in the sense that DX12 is comparable to it.
4970K runs at 4.2 stock. Plz.
It would be interesting to see Mantle alive and well just not to let Microsoft sit on their ass. Just ponder over the DX11 situation : it's been 5 years since it has been released and developpers and IHVs have been pressuring MS to release a more lightweight API. Yet, DX12 will only release this fall. What makes Mantle attractive from a consumer standpoint is that it competes directly with MS's DX so they could feel the need to update DX12 more regularly.
It's attractive because it's competition, even if only compatible with GCN hardware. I agree with you that I don't see it gaining any more traction after DX12 is released but I would certainly like it to. With regards to supported games Mantle had more games in one year than Directx 11. Not bad.
Well I won't complain if it leads to a more competitive place. Otherwise chances are we will be stuck with DX12 until 2020.
