Increased number of frames and computational complexity of focus measure causes Shape From Focus (SFF) method to become memory hungry and time consuming technique. These factors become limitation for SFF techniques to be used in fast and real time applications. However, computational time can be reduced significantly with parallel implementation of these methods on multiple cores. In this article, focus has been measured through Discrete Cosine Transform (DCT) and speedup performance has been tested at different number of cores; as a result of which one can easily decide about the optimal number of cores required for SFF application. © 2013 American Scientific Publishers All rights reserved. All rights reserved.