If you are blessed with the good fortune of already having a VR headset and happen to be running an NVIDIA GPU then there is a new driver you want to grab as soon as you can. The driver includes a new OpenGL extension that enables NVIDIA SLI support for OpenGL apps that display on an Oculus or Vive. NVIDIA's PR suggests you can expect your performance to improve 1.7 times, not quite doubling but certainly offering a noticeable performance improvement. The update is for both GeForce and Quadro cards.
They describe how the update will generate images in their blog post here. They imply that the changes to your code in order to benefit from this update will be minimal and it will also reduce the CPU overhead required to display the images for the right and left eye. Read on if you are interested in the developer side of this update, otherwise download your new driver and keep an eye out for application updates that enable support for SLI in VR.
Nvidia says that this driver
Nvidia says that this driver fixes the 144Hz issue with GPUs remaining at higher than idle clockspeeds.
In my testing this issue has been only partially fixed. I have a single Jetstream 980 Ti. I have two 27″ 2560×1440 monitors. First is an ROG Swift, 144Hz G-Sync display. The second is a Dell U2711, 60Hz.
With the Swift alone connected, the issue has been fixed. Before this driver at 144Hz at the desktop I’d idle at 925MHz. With the new driver I idle at 135MHz. The previous solution was to run the monitor at 120Hz at the desktop, that led to the card idling correctly.
With both monitors active, I still get the 925MHz idle clockspeed if the Swift is at 144Hz at the desktop. If I lower it to 120Hz with the Dell at a constant 60Hz, then I once again idle at 135MHz.
It seems to be an issue with multiple displays too. 🙁 Perhaps some further testing is required on Nvidia’s part? Good to see it fixed for at least single display setups.
It won’t reduce the CPU
It won’t reduce the CPU overhead required compared to AMD’s GPU hardware, if the Nvidia GPU hardware does not have Fully in hardware asynchronous compute ability and fully in hardware GPU processor thread dispatch and scheduling ability. The CPU is going to have to run Nvidia’s software based GPU processor thread dispatch and scheduling in software, and not be able to fully keep Nvidia’s GPU execution resources from sitting idle for lack of the much faster and fully IN hardware GPU processor thread dispatch and scheduling ability of AMD’s GCN graphics/compute.
There are no software based processing CPU/GPU/Other thread dispatch and scheduling abilities that are as fast and responsive, and able to keep those CPU/GPU/Other processor threads and processor execution resources as efficiently utilized as the fully in the hardware types of thread dispatch and scheduling abilities. Software based scheduling is just too slow, and results in the hardware execution resources remaining idle for lack of efficient CPU/GPU/Other processor thread dispatch and scheduling abilities.
A tad too late Im afraid.
A tad too late Im afraid. Maybe next time.
By the time VR is actually
By the time VR is actually here and mature we will have new Nvidia GPU with all this in hardware so your arguments are moot.
We still dont have anything DX12 final either. Sit and wait some more pls….
It better be, or Nvidia will
It better be, or Nvidia will have problems! I do not see Nvidia getting all the asynchronous compute fully in its GPU’s hardware done and certified to work in one generation, Nvidia has so Gimped its hardware of compute features. Maybe Nvidia went a little too far in trying to do things with the software and forget about improving the compute side of things in its hardware, and that includes asynchronous compute supported fully in its GPU’s hardware.
Nvidia was sure stripping out that compute and trying to sale users on the power savings at the expense of future VR performance! Nvidia sure hired AMD’s Phil Rogers(AMD’s HSA/Fellow and HSA Foundation representative) quickly, but that radical of a change in GPU design philosophy from Nvidia will not be completely done by Pascal’s release date.
Nvidia still have sometime
Nvidia still have sometime until pascal release. They still have plenty of time to tweak their architecture. Back in 2009 they say nvidia will never beat AMD in tessellation because AMD was mile ahead of nvidia when it comes to hardware tessellation (AMD have the hardware ever since tessellation still not part of DirectX). But when nvidia finally comes out with their tessellation engine they easily much better than AMD. And till this day AMD still behind nvidia when it comes to tessellation.
At least they got bitten in
At least they got bitten in the ass and are woken up by it now so expect something coming from them on this. 2016 nah more like 2017+