Could you link to the documentation and test of this?
Really, what in the entire gaming industry history made you think that?
It's better than free because it reduces latency too. This is why GCN is so much better than Maxwell in VR.
If they were really sponsored by AMD I doubt this would have happened.
I dunno for YOUR case. But for Lionhead's case, they seem to think Async Compute is "free performance" for GCN.
I take their words over forum warriors with no experience in DX12 programming, any day of the week.
If they were really sponsored by AMD I doubt this would have happened.
You should re-read the quote again. Nowhere does he blame, he said it could be a MB issue.
Starswarm is the perfect example of Oxide's neutrality.
You mean Star Swarm that on purpose ran slower than it should on nVidia hardware at release?
Could you link to the documentation and test of this?
Despite the polished and impressive graphics, which represented some of the best Ive seen anywhere in VR, the entire demo ran flawlessly on the Oculus Rift Crescent Bay prototype, all at 90Hz.
Kite & Lightning co-founder Ikrima Elhassan told me that the GE Neuro demo is powered by a single AMD R9 Fury X GPU, formerly codenamed Fiji. The team employed AMDs new LiquidVR system which Elhassan said was critical to achieving the performance seen in the demo.
Without late latching/Liquid VR, we were seeing GPU idle times of 40%! he told me.
http://www.roadtovr.com/hands-on-am...e-lightnings-spectacular-ge-neuro-experience/
I suppose it's possible that GE are being sponsored by AMD as well though.
Nope, the Star Swarm that improved massively on NV performance, without Oxide crippling NV's DX12, which was later used by some forum warriors to boast how awesome NV is at DX12 over AMD.
Oxide's actions have shown them to be neutral. No need to drag them into the mud for your agendas.
If you want to bash some devs, turn your attention to Ubisoft & Warner Bros.
Btw, here's some info for you guys, to show if Oxide wanted to cripple DX12 for NV, they could easily do so:
Also on there, is the statement that DX12 code examples is almost identical to Vulkan/Mantle, for anyone else doubting the origins of these new APIs.
Wow. I dont even now what to say...
DX12 allows much better optimization on GPUs than DX11. It has nothing to do with nVidia or AMD. The whole API is better suited to new GPU architectures than DX11. That's the reason why it is even supported on Fermi - a GPU architecture from 2010, introduced 6 months after DX11...
You should re-read again what everybody is saying about these new low level APIs starting with the AZDO from OpenGL 4.4.
In the end you just said, that DX12 is bad for nVidia because DX12 is able to do more on their hardware. :|
I asked for documentation and tests.
And you got something better, a real world example. You can find plenty of documentation on LiquidVR if you want, it's all over the web.
I can find just the same with Gameworks VR being better.
Where?
Games like Elite.
But again, its pointless until someone can actually document it.
Elite doesn't use GameWorks VR. GameWorks VR only came out of alpha a week or two ago.
Yet already supported by a few companies in beta versions.
What games uses LiquidVR? 0?
But again i asked you for something you couldnt deliver.
Have there been any gamedevs from major studios that have praised NV's GPUs in relation to DX12 async compute? I'm trying to google some examples but can't find any.
I've looked quite hard but my search has been fruitless. Remember I'm looking for evidence from game developers themselves, not from Nvidia.
It's better than free because it reduces latency too. This is why GCN is so much better than Maxwell in VR.
Because you are searching for AMD terms?
I dont find any Intel CPUs if I search for APUs either.
Maybe I need to remind you of your original claim,
You still havent shown any data on this.
Asynchronous shading is an incredibly powerful feature that increases performance by more efficiently using the GPU hardware while simultaneously reducing latency. In the context of VR, it is absolutely essential to reducing motion-to-photon latency, and is also beneficial for conventional rendering.
To avoid simulation sickness in the general population, the motion-to-photon latency should never exceed 20 milliseconds (ms).
20 milliseconds or less will provide the minimum level of latency deemed acceptable.
Typically, it takes 50 milliseconds or more to render a scene in a virtual reality environment, from the time you move your head to the time you see a response. Maxwell cuts that latency in half.
The standard VR pipeline from input in (when you move your head) to photons out (when you see the action occur in-game) is about 57 milliseconds (ms). However, for a good VR experience, this latency should be under 20ms. Presently, a large portion of this is the time it takes the GPU to render the scene and the time it takes to display it on the headset (about 26ms). To reduce this latency we've reduced the number of frames rendered in advance from four to one, removing up to 33ms of latency, and are nearing completion of Asynchronous Warp, a technology that significantly improves head tracking latency, ensuring the delay between your head moving and the result being rendered is unnoticeable.
Combined, and with the addition of further NVIDIA-developer tweaks, the VR pipeline is now only 25ms.
NVIDIA suggests that VR Direct will reduce latency "on average" by 10 milliseconds - 10ms from the OS and 4ms from the GPU, also on average.
Asynchronous shading is an incredibly powerful feature that increases performance by more efficiently using the GPU hardware while simultaneously reducing latency. In the context of VR, it is absolutely essential to reducing motion-to-photon latency, and is also beneficial for conventional rendering.
Excellent post @VR Enthusiast. Backed by solid evidence from people in the know.
I had heard devs stated that GCN is excellent for VR due to the latency (to prevent motion sickness), the oft remark is that Maxwell was less capable, but I did not realize it was *THAT* bad.
Any thoughts why Kepler/Maxwell struggles with VR?
Because this implies they can't do Async Shading very well:
On Maxwell 2 the compute is not stateless, so a context switch might be needed to perform an async task, and this comes with a huge performance penalty. The GCN is a stateless compute hardware, so it can run any async compute task without context switch.
Pascal won't magically solve the problems for NV, because they don't have an own API like Mantle.
Originally Posted by Silverforce11: One of the features for VR to reduce latency is also called Async Warp.
You misunderstood what I said. For DirectX 12 you have only API driver that talks to the game engine and to the hardware. In DirectX 11 you have had API and a driver for the gpu. Thats where you, if you were GPU vendor, optimize the hardware and software for specific game and scenario. With DirectX 12 you can't.
In essence we can put this this way. The API is the driver for the hardware. You have no room to optimize the game-hardware path because the API talks directly to the hardware. It is only possiblr with optimizing the API driver. With its multithreaded capibilities on AMD GPU it is bottlenecked with rasterizing performance. With its rasterizing demand its bottlenecked on asynchronous compute on NVidia GPUs.
The biggest problem is that the game engine works ONLY with the API. The API talks directly to the hardware. It is due your hardware to be able to use the best as it is possiblr of the capabilities the API offers. If you hardware is not able to get enough work scheduled it will be bottlenecked. If your GPU is not able to get enough pixels pushed it will be bottlenecked.
We can argue abiut this. Blaming Oxide is like blaming Microsoft for giving DirectX 12. With this API you can't lock any developer from performance. It is simply not possible. You cannot make one vendor look better than the other. That is not the nature of DX12. And no, im not software engineer. But i understand what means that the application talks directly to the GPU, without the CPU intervention.
I can see your point that malicious developers can purposely hinder a vendor.This kind of thing is much easier to do in DX12 vs. DX11. You can simply send the GPU commands that work well on one set of hardware but not on another, and due to the lack of driver overhead, the driver cannot correct for this.
As a handwaving explanation consider programming in Java (DX11) and C++ (DX12). A well written C++ program should always be faster than the java equivalent. However java can be run with some very aggressive performance enhancements and libraries (driver intervention) so that a poorly written C++ program will run worse than that same program in java as the JRE can add performance enhancements. As well in a program written in C++ your performance optimizations can favour one architecture over another (ie loop unrolling).
Its the same thing with DX12 and Oxide (hate them or love them) have stated as much.