Pemrosesan Paralel pada Low Pass Filtering Menggunakan Transform Cosinus di MPI (Message Passing Interface)

Abstract

Parallel processing is a process of calculating two or more tasks simultaneously through the optimization of the computer system resource, one treatment models is a desktop system. The model allows to perform parallel processing between computers with specifications different computers. An implementation of a model network of workstations using MPI (Message Passing Interface). In this study, applied to the case of the low-pass filtering (LPF), a process in the image or the image of the shape of the filter that retrieves the data at low frequencies. filtering programs lowpass using the cosine transform MPI implemented by modifying the algorithm in the process on each node (computer). Depending on the test results, so that the processing speed of a parallel system is influenced by the number of nodes / processes and the number of frequency components that are processed. In the treatment of single larger process, the time it takes more and more and the value prop affects only the amount of high frequency data is filtered on the field. While parallel processing of more and more computers involved in the filter calculation process low-pass, plus the time required to perform the calculation.