GTC 2012: NVIDIA Announces GeForce GRID Cloud Gaming Platform
NVIDIA puts its head in the clouds
Today at the 2012 NVIDIA GPU Technology Conference (GTC), NVIDIA took the wraps off a new cloud gaming technology that promises to reduce latency and improve the quality of streaming gaming using the power of NVIDIA GPUs. Dubbed GeForce GRID, NVIDIA is offering the technology to online services like Gaikai and OTOY.
The goal of GRID is to bring the promise of "console quality" gaming to every device a user has. The term "console quality" is kind of important here as NVIDIA is trying desperately to not upset all the PC gamers that purchase high-margin GeForce products. The goal of GRID is pretty simple though and should be seen as an evolution of the online streaming gaming that we have covered in the past–like OnLive. Being able to play high quality games on your TV, your computer, your tablet or even your phone without the need for high-performance and power hungry graphics processors through streaming services is what many believe the future of gaming is all about.
GRID starts with the Kepler GPU - what NVIDIA is now dubbing the first "cloud GPU" - that has the capability to virtualize graphics processing while being power efficient. The inclusion of a hardware fixed-function video encoder is important as well as it will aid in the process of compressing images that are delivered over the Internet by the streaming gaming service.
This diagram shows us how the Kepler GPU handles and accelerates the processing required for online gaming services. On the server side, the necessary process for an image to find its way to the user is more than just a simple render to a frame buffer. In current cloud gaming scenarios the frame buffer would have to be copied to the main system memory, compressed on the CPU and then sent via the network connection. With NVIDIA's GRID technology that capture and compression happens on the GPU memory and thus can be on its way to the gamer faster.
The results are H.264 streams that are compressed quickly and efficiently to be sent out over the network and return to the end user on whatever device they are using.
On the client side, GeForce GRID comes into play again to decompress the H.264 stream and then render the final image to be displayed on the device. While this will obviously work on the PC and laptop side of things, there are still questions on whether GRID technology is necessary on the consumer end of phones, tablets and even TVs. Will NVIDIA require Tegra-based tablets and smart phones in order to take advantage of the streaming services with GRID–and does that mean that we will see NVIDIA-powered TVs in the near future?
UPDATE: The answer is no, it does NOT require NVIDIA technology on the decode side and in fact NVIDIA CEO Jen-Hsen Huang said "just about any decent H.264 decoder will work."
The primary benefit of GeForce GRID technology is that it is supposed to lower the latency between rendered images and the time the user sees them–and how soon they can respond to them. The lower bar on this graph, provided by NVIDIA, represents what a typical user's latency would be on a "console gaming system." The times here are pretty vague, though the 100 ms game pipeline and the 66 ms display latency seem somewhat reasonable.
The middle bar represents the first generation of streaming gaming - including services like OnLive - that I have discussed (and berated) in the past. Total latency in this case has gone from 166 ms from game time to display time all the way up to 286 ms with addition of 30 ms of capture and encoding time, 75 ms of network latency and 15 ms of decode latency on the client side.
NVIDIA GRID aims to bring the total latency UNDER that of current generation consoles. You still have latencies for capture, network and decode, though they are noticeably lowered. Game pipeline time is cut in half thanks to the performance of Kepler GPUs, and with the GPUs ability to quickly pass information from the frame buffer to the encoders. Capture and encode has gone from 30 ms down to 10 ms with the fixed function NVENC (hardware video encoder) unit on the Kepler graphics units. Decode time is also decreased on the client side with NVIDIA technology and even network latency is reduced from 75 ms to 30 ms.
Honestly, the only timing reduction I have a problem with is the network latency–how NVIDIA plans to solve the problems of physics involved with network infrastructures needs to be proven. The idea of "estimations" or "predictions" can work with some game types better than others, but I think hardcore gamers are going to skeptical until proven otherwise. The only real answer is to build out more data centers to make sure there is one withint a reasonable distance to the consumer.