Hi,
I have good knowledge on CUDA, I can handle CUDA projects efficiently.
There is an important point that many peoples ignore, CUDA will not always give better performance than the CPU, the only way to get full power of GPU is that the formulas or algorithms, can be fully parallelized, with no inter-communication between threads otherwise you will have to use CUDA atomics operations which are sequentials.
Best Regards,
Marouane