How Nvidia is pursuing the 'holy grail' of real time photorealistic VR
- 31 March, 2018 08:40
Nvidia continued to tout its credentials as the underlying technology for designing and deploying photorealistic virtual reality (VR) systems during its GTC Technology conference in Silicon Valley this week, including the announcement of its most graphics card to date and a new system specifically designed for testing autonomous driving tech.
Cand founder Jensen Huang reasserted that real-time rendering of photorealistic images in VR is "the holy grail" for the industry and asserted that: "Recreating virtual reality is one of the most daunting computing tasks we know and yet, on the other hand, it is an enormous industry."
Nvidia's latest proposed solution combines both hardware and software with the Quadro GV100 graphics card and RTX ray tracing technologies, with the aim of achieving the elusive goal of delivering intensive ray tracing imaging in real time.
Ray tracing is a rendering technique which traces the path of light rays in an image as they illuminate a scene and bounce off objects, which is vital to creating photorealistic VR experiences and would be nothing short of groundbreaking for the gaming industry.
This is particularly suited for the design of compute-intensive VR content, giving designers and developers the opportunity to identify high-quality lighting and shadows in imagery.
The combination of the Quadro GV100 and RTX has also been built to reduce the waiting time for developers, adding the ability to perform rendering up to 10 times faster than the traditional CPU.
Speaking at the chip-maker's annual GTC Technology conference this week, Huang said: "With the Nvidia RTX Technology and the Quadro GV100, I believe the number of frames that will be rendered will jump by a factor of 10, just so that you can iterate and try things before you get it done.
"You’ll also get it done faster but most importantly, you'll save money. The more GPUs you buy, the more money you save."
"After 10 years, what makes this special is for the very first time we can bring real-time ray tracing to the market," he added. "People can actually use it. Technology has been encapsulated into multiple layers from our GPU architecture to the algorithms that makes it possible for us to do this, you are also seeing deep learning in action."
What is the Quadro GV100 and RTX?
Huang unveiled the new workstation-based graphics card, the Quadro GV100, at GTC this week, with the company claiming it to be the most powerful graphics card released to date, knocking the Titan V down the pecking order.
It comes with 32GB of memory, which can be scaled up to 64GB with multiple Quadro GPUs by using NVLink . The technology is based on Nvidia Volta GPU and contains 7.4 teraflops of double precision (64-bit) and 14.8 teraflops of single precision (32-bit), all combined with 118.5 teraflops of deep learning performance.
Nvidia's Quadro GV100 is the first GPU to be built on the Volta architecture. It has powered Oscar nominees throughout the last 10 years and acts as a visual computing platform.
Huang explained that the technology includes the ability for developers to access Nvidia's RTX ray tracing technology through its OptiX application programming interface.
GPU developers are also able to access its new DirectX ray tracing API using Nvidia's Quadro GV100 and RTX technology thanks to a partnership with Microsoft.
The Quadro GV100 is available now for purchase on the company's website at $9,000.
Earlier this week the company unveiled that it had doubled the memory of the Tesla V100 GPU to 32GB in order to tackle the most memory intensive deep learning and high-performance computing workloads.