The New Era Of GPUs And Graphics Rendering

18th July, 2019 by

GPU rendering technology has been around for years, capitalizing on the high-speed graphical rendering which modern processors are capable of. Graphics processing units are now powering an exciting new era in rendering, displaying computer graphics with a level of realism that simply has to be seen to be believed.

It’s worth taking a moment to consider the difference between dedicated GPUs and their CPU counterparts. Central Processing Units (CPUs) are a computer system’s general-purpose brain, responsible for the computing power behind many tasks and processes. Traditional CPUs had a single core design and were able to perform one task on one bit of data at a time. Today, most CPUs have somewhere between six and 14 cores. They’re able to run 12 to 28 threads of instructions simultaneously, with each thread handling one block of data.

On-screen relationship

The primary task of a GPU is to process visual data for outputting on-screen. In order to be as effective as possible, GPUs place an emphasis on parallel processing. They’re designed to run instructions across multiple cores on multiple blocks of data; each GPU can have thousands of cores, and run over a hundred threads of instructions. Each thread can also work on 30 or more blocks of data at once. As a result, a GPU might process around three thousand blocks of data at a time, while a CPU could only process around 24 blocks.

This huge difference in performance explains the correlation between GPUs and graphics rendering. Consider the rendering of one HD frame composed of two million pixels. The difference between a CPU and a GPU is the difference between processing 24 of those pixels at a time, or 3,000 at once.

Ray of light

The relationship between GPUs and graphics rendering is not new, but the development of ray tracing certainly is. Some experts believe that ray tracing represents the biggest leap forward in computer graphics for decades – and it’s all being made possible by the next generation of high-performance GPUs.

Ray tracing is a rendering technique which generates an image by tracing the path of light as pixels in an image plane. The technique is then able to simulate the effects of the light rays’ encounters with virtual objects. Simulated effects include reflections, refractions, and scattering. meaning that ray tracing can introduce ultra-realistic reflections, shadows, and other effects into graphics, producing an extremely high degree of visual realism.

The power and the glory

However, ray tracing requires huge amounts of processing power. Consequently, it was formerly limited to applications where frames didn’t need to be rendered quickly – film and TV visual effects, plus still image processing. The newest GPUs make ray tracing applicable to real-time applications like computer games, performing several new tasks to quickly render ray-traced effects. These include a process called Bounding Volume Hierarchy, or BVH, and BVH determines which aspects in a game scene any ray will intersect with. GPUs can also apply a denoising algorithm, to improve an image’s visual quality while reducing the number of rays needing to be cast.

The end result is graphics that deliver new heights of attractiveness and realism. Developers are already starting to recognize the potential of GPU-powered ray tracing, and many big game franchises are planning to leverage this technology for their newest titles. Sony has made a point of revealing its much-anticipated PS5 console will support this rendering technique. There’s little doubt that when it comes to GPUs and graphics rendering, ray tracing represents the future