folder

I think it would need to recoded the whole program to push GPU to calculate it. It's really hard task, with CUDA support of Nvidia, it would be easier but still hard. It's better to use more threads for now on.