The only downside is that it's been written in CUDA which only runs on Geforce 8 cards are newer. It might have been better to write the code in either openGL or even DirectX, so that everybody could take advantage of faster rendering. But I guess there were good reasons not to.
