Encoding & Rendering Benchmarks

Though the primary focus of this article is on real world performance, synthetic benchmarks allow us to set some baseline performance numbers across a variety of possible usage scenarios. This allows you to compare our results to that of other setups that have done these same synthetic benchmarks. We were unfortunately limited by the GTX 9800 GX2 in some of the benchmarks due to it only being a DirectX 10 capable GPU. Despite not being able to use DirectX 11 benchmarks, we wanted to continue with the 9800 GX2 because it came out at approximately the same time as Skulltrail, and we wanted the hardware to belong to the same time period.




The NUC just barely beats out the Skulltrail setup in the CPU category; however, Cinebench scales quite well with cores and frequency, so this should be expected given the number of cores/threads both setups have, and there is the similar frequency of the CPUs to consider.




On the OpenGL side of the story, the integrated GPU on the NUC really starts to show what it is capable of, nearly doubling the scores in both the R15 and R11.5 benchmarks.



Next up is the x265 codec benchmark, which is generally considered a much more "real world" scenario test that encodes a 1080p clip into x265 format. The results here were not so close as they were with Cinebench. Despite the NUC's size, it manages to encode the clip around 40% faster (76 seconds on average as opposed to the 110 seconds with Skulltrail) than the Skulltrail setup. This speaks very positively for those looking to use the NUC to build mini video production workstations, either for startups or for crews that are going on the road.



We also ran the Dolphin benchmark, which is a small test that checks on how well your system can run the now infamous Dolphin emulation program for Game Cube and Wii - more infohere.