Ryzen 9 3900X/3950X vs. Core i9 10900K In 380+ Benchmarks
When looking at Intel's own software offerings that are part of their oneAPI toolkit including the likes of Embree and Open Image Denoise and then the Embree-using Tungsten renderer, the i9-10900K was faster than the 3900X but the Ryzen 9 3950X offered significantly better performance.
With the Python workloads from PyPerformance, PyBench, and Numenta with them being predominantly single-threaded, the Core i9 10900K came out in the lead due to being able to clock higher with its turbo frequencies.
Meanwhile for thread-happy ray-tracing workloads with C-Ray, Tachyon, rays1bench, and YafaRay, the Ryzen 9 3950X was 7.6% faster than the i9-10900K.
Or if expanding the scope to include all of the renderers tested with the likes as well of Blender, Appleseed, LuxCore, TTSIOD, IndigoBench, the 3900X was about 4% faster while the 3950X was 34% faster than the Intel Comet Lake CPU under test.
When looking at the video encoding tests with FFmpeg, x264, x265, libvpx, and AOM AV1, the i9-10900K did mange to secure another victory.
Lastly is a look at the workstation class workloads where the Ryzen 9 3900X is 11% faster than the 10900K and the 3950X has a 21% advantage.