NVIDIA is among the world’s largest graphics card manufacturers. Thanks to the technologies and performance graphics cards it offers to its users, the company is loved and bought by many game lovers.
The company uses RTX (ray tracing) technology to offer its users better lighting and shading in games. NVIDIA is working on a new technique for ray tracing technology, according to a new article it published. Allegedly, thanks to its new technology, new graphics cards will offer better ray tracing technology than currently.
NVIDIA will offer 20 percent better RTX technology in the coming years!
According to the information revealed by Twitter user Ox22h, NVIDIA recently published an article explaining the new development. In this article, the company talks about a new technology, Subwarp Interleaving. Thanks to this technology, ray tracing technology is expected to offer 20 percent better performance on NVIDIA graphics cards that will be introduced in the coming years.
GPU Subwarp Interleaving works by better optimizing the elements of ray tracing technology. In normal use, it offers a performance increase of 6.3 percent compared to the current technique. This increase is as high as 20 percent in some places. However, this system does not seem very suitable for the currently used architecture. It needs to be done with a change in the architecture for the performance increase.
In the article published by NVIDIA engineers, the currently working RTX technology and the problem were mentioned:
“First, GPUs group them into units we call warps, which are taken from a single program counter (PC) and executed in a SIMT (single instruction, multi-thread) fashion. Second, GPUs hide pauses by scheduling simultaneously between many active warps. When a long latency operation stops a warp and the GPU’s warp timer cannot find an active warp to bypass, a subwarp timer may instead switch execution to another different sub-warp of the current warp. ”
The new technique Subwarp Interleaving reduces distortion latency on GPUs, giving a better result in such situations. This technology will most likely appear in GPUs after the RTX 40 series. For this technology, the company will produce a new architecture. So, what are your thoughts? Do not forget to give your comments and opinions below.