Running this on GPU will be mush faster,. Let me see if i can write a CUDA program for this.