NVIDIA + mental image = was?

Recently, many 3D sites have reported that NVIDIA has acquired mental image. I think this is a strong signal for both the film industry and the desktop PC graphics industry.

After NVIDIA acquired Larry gritz and renderman's brother intropy, it immediately launched gelato, which did not change its display window, it is known as a movie-level Renderer that can be accelerated with its own GPU. Just a few weeks ago, Intel acquired neoptica, a company organized by a strong Stanford team. According to industry speculation, the back-end may be NVIDIA. The company neoptica was very interesting because these people were not satisfied with the current GPU design. Several papers were made by neoptica. Some people even predicted that when NVIDIA acquired mental image, real-time line tracking on the GPU will be available in the future. I think this is a bit exaggerated, because the current interactive ray tracing technology is still mastered and respected by Germans, and its representative character is, j örn schmittler, which has been developed by Intel, Ingo Wald, which is now in Utah, and a group of researchers at Max Planck sort ut. NVIDIA is a company that has lived on Raster hardware for more than a decade. It has no strength in the field of ray tracing to compete with Intel, lack hardware support, and lack of developers. From a technical point of view, NVIDIA will be an important competitor in the future, rather than AMD/Ati. AMD has a gray face everywhere. The CPU won't work for Intel, and the GPU won't work for NVIDIA. I really don't know what AMD's fate will be like in the future.

NVIDIA's cleverness lies in its rich product lines and sound developmentCommunitySupport and friendly software development, which is similar to Microsoft. Tesla has just obtained magazine technical excellence award. NVIDIA must be very proud of it because GPU stream technology has received support from the grassroots. So far, NVIDIA may change its product strategy in the next few years. In order to deal with the surging intel and the active AMD/ATI, it will gradually enter the offline rendering market that has previously been involved in depth. Customers will not only include volkswage, Boeing, but also film studios such as ILM. According to ILM, only when renderman is insufficient can Mental Ray be used for fire relief. renderman is purely designed for the film industry, while Mental Ray includes not only movies but industrial customers. Renderman's advantage lies in speed and image quality. It supports a wide range of academic and application fields. The most important thing is whether renderman is a 90% or raster. Mental Ray is the core of high-performance raytracing and uses ray tracing to solve all problems. Now NVIDIA has the menta ray technology, and it is very likely that it will enrich its own gelato and map it to the Tesla platform. This not only maintains good compatibility, but also solves the huge problem of shading performance. It can be said that there are still many things to be done, especially software. I hope to improve the streaming programming conditions in the future, write c ++ meta as rapidmine, and then generate glslCodeThe method is too limited. Cuda and staticProgramThe coupling is too tight. The virtual machine Hal of the CTM does not know the performance, and the use of Brook is still cumbersome. Otherwise, in the future, just get a stream Vm, dynamically execute it under the condition of guaranteed performance, and map a lot of OP to the GPU?

The CG industry is full of mountains and rain. Let's wait and see.