Parallel bucket sorting on graphics processing units based on convex optimization


Autoria(s): Beliakov, Gleb; Li, Gang; Liu, Shaowu
Data(s)

01/01/2015

Resumo

We found an interesting relation between convex optimization and sorting problem. We present a parallel algorithm to compute multiple order statistics of the data by minimizing a number of related convex functions. The computed order statistics serve as splitters that group the data into buckets suitable for parallel bitonic sorting. This led us to a parallel bucket sort algorithm, which we implemented for many-core architecture of graphics processing units (GPUs). The proposed sorting method is competitive to the state-of-the-art GPU sorting algorithms and is superior to most of them for long sorting keys.

Identificador

http://hdl.handle.net/10536/DRO/DU:30056837

Idioma(s)

eng

Publicador

Taylor & Francis

Relação

http://dro.deakin.edu.au/eserv/DU:30056837/beliakov-parallelbucket-2015.pdf

http://dro.deakin.edu.au/eserv/DU:30056837/beliakov-parallelbucket-inpress-2013.pdf

http://dx.doi.org/10.1080/02331934.2013.836645

Direitos

2013, Taylor & Francis

Palavras-Chave #convex optimization #cutting plane method #GPU #order statistic #parallel selection #parallel sorting
Tipo

Journal Article