Time reduction in video compression using the SAD algorithm in CUDA

This paper gives a new insight into the computation of the Sum of Absolute Differences Algorithm for Video Compression. We implemented the SAD algorithm using various techniques in CUDA to improve the time taken for computing the SAD value. These techniques were then compared to the MATLAB simulated code and the conventional CPU code. We compared all the computational times and concluded that CUDA techniques gave a new exemplary dimension to the field of computing. In the near future, millions of processors could be used for computing at once transforming computation speed from seconds to nanoseconds.