Performance Analysis of FDCT Algorithms

The book identifies 4 popular FDCT algorithms. Next it is mapped in a cluster architecture.The cluster specific parallel algorithm is written. The algorithms are analyzed through a timing diagram and the cluster utilization and time units taken to implement the algorithm is calculated.Proposed algorithm parallel DCT is introduced and explained.It is also mapped in the cluster architecture, cluster specific algorithm is written,timing diagram is constructed, and cluster utilization and time units taken is calculated.It is shown thorough graphs and tables that cluster utilization is highest and time units taken is lowest in parallelDCT. It is concluded that, in order to minimize the number of multiplications in popular FDCT algorithms cluster utilization is ignored. Whereas parallelDCT contains more number of multiplications but the cluster utilization is more and time unit is less. Hence the simple approach of parallelDCT is more effective.

Autorentext

Atri Sanyal is an Assistant Professor of Department of Computer Application of NSHM College of Management & Technology, Kolkata. He did M.Sc(CS) from Banaras Hindu University and later on did ME(CSE) from West Bengal University of Technology.His research areas are Computer Architecture, Image Processing, Parallel Architecture, Parallel Algorithm.