This Sounds Good for XPlane 10 Future re: GPU's

It's not just tesselation itself that will gain performance wise, its also the fact that in asking the GPU to attempt tesselation on the mesh, it also must render the final product, with minimal overhead of the CPU, freeing it up for other tasks. Essentially, this helps to offload the mesh rendering to the GPU and using some of the specific capabilities of the GPU directly to improve a low-res mesh, but without the other items like a bump map and normal map and other things, will only marginally improve, but still the performance gains of CPU offloading can help bump those sliders right or make room for features not yet in place that they're waiting on performance to be optimal before implementing.