With two RTX 5000 cards using NVIDIA NVLink for a combined 32 GB of memory, we have been able to ray trace scenes containing 1.37 billion unique triangles. The GPU can handle pretty complex scenes though. It is possible to swap textures from CPU memory to the GPU. Very complex scenes with a lot of geometry and textures may not fit on the GPU, which leaves the CPU as the only choice. One of the challenges with GPUs is running out of memory. Once the data is uploaded and shaders are compiled, the workflow is very smooth and fast. KeyShot runs very fast using the new NVIDIA RTX cards. However, the large number of compute threads on the GPU does allow for much higher throughput and the recent addition of fast denoising algorithms has further bridged the gap between brute force GPU algorithms and more sophisticated CPU algorithms. This does mean that the GPU algorithms converge more slowly to a noise-free image than the CPU algorithms. The GPU rendering is using slightly different algorithms as GPUs perform best with uniform parallel workloads.

As a result, Keyshot users will be satisfied with CPU server performance at iRender. For CPU rendering, iRender also provides CPU servers with unmatched speed for rendering Keyshot with Dual Xeon E5-2670 v2 2.50 GHZ, 20 cores, 40 threads, RAM 128 GB, Storage 256 GB and a Geforce GPU Nvidia 1050 2GB, which is totally higher than recommended system requirements. In 2018, this all changed when NVIDIA introduced the RTX architecture with dedicated ray tracing hardware. The ray tracing code at the time was running on the AMD Opteron architecture. Luxion demonstrated interactive ray tracing technology to the public in March 2006.

Zbrush keyshot render