P-TWDTW: Parallel Processing of Time Series Remote Sensing Images Using Manycore Architectures
ResumoIn the class of computationally complex problems, the time series analysis is one of those that has high demand for computational power, due to the complexity of the algorithms and the large volume of data to be analysed. The TWDTW algorithm stands out as of the best solution found in the literature in this field, but its time complexity O(n2) makes its unfeasible for large data sets. This work proposes a parallel algorithm, called P-TWDTW (Parallel TWDTW), that allows analyzing large scale time series exploring Manycore (GPU) architectures. In the evaluation of the algorithm, the P-TWDTW proved to be a promising solution with response time up to 11 times lower than TWDTW.
Palavras-chave: Graphics processing units, Random access memory, Time series analysis, Informatics, Parallel processing, Computer architecture, High performance computing, TWDTW, Parallel Processing, GPU, Time Series
DE OLIVEIRA, Sávio Salvarino Teles; RODRIGUES, Vagner J. do S.; FERREIRA, Laerte G.; MARTINS, Wellington S.. P-TWDTW: Parallel Processing of Time Series Remote Sensing Images Using Manycore Architectures. In: SIMPÓSIO EM SISTEMAS COMPUTACIONAIS DE ALTO DESEMPENHO (WSCAD), 19. , 2018, São Paulo. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2018 . p. 252-258.