Researchers find quantum computing can improve ray tracing performance by 190%
Researchers find quantum computing can improve ray tracing performance by 190%.
Researchers in the US, Portugal and the UK predict that the solution to ray tracing’s high performance requirements may be to mix older ray tracing algorithms with quantum computing.
In a recently published research white paper, quantum computing enhanced ray-tracing workloads with a 190 percent increase in performance.
This process is done by limiting the number of computations required per ray.
Ray tracing in graphics technology has made an evolutionary leap in games, especially how games are rendered.
However, performance and a developer’s ability to properly adopt the process are secondary to complexity.
The problem lies in the hardware and computational requirements of ray tracing technology, and the necessity for specific hardware, which limits the use of the core technology for most users.
The researchers describe how quantum computing has the potential to minimize the processing tax caused by ray tracing technology.
The team took a 128×128 image with ray tracing enabled and optimized it using three different strategies.
The three passes are classical rendering techniques, unoptimized quantum rendering, and then optimized quantum rendering.
The first technique computes 2.678 billion ray intersections on the 3D image, giving 64 for each individual ray.
The unoptimized method halved the first number and required only 33.6 ray intersections, which is equivalent to 1.366 billion ray intersections.
Using optimized quantum techniques and classical systems, the final attempt presented an image of 896,000 intersections with 22.1 intersections per ray.
The most important disadvantage of the technology is the quantum computing system.
Quantum computers and devices are currently developing NISQ, the Noise Intermediate Scale category of quantum products.
These intricate systems are not the most performant, so rendering takes hours to calculate each image correctly.
This category is great for simulation, but currently it’s hardly a viable option for rendering games.
While the results are good, the technology is far from being ready for production.
In the past year or two, with current quantum computing trends, we have only seen a small amount of quantum computing available.
IBM plans to increase the amount of quantum computing in the next few years, but how much progress the technology will have in a short period of time is unknown.