923

Fast parallel GPU-sorting using a hybrid algorithm

E. Sintorn, U. Assarsson
Department of Computer Science and Engineering, Chalmers University Of Technology, Gothenburg, Sweden
Journal of Parallel and Distributed Computing, Vol. 68, No. 10. (October 2008), pp. 1381-1388.

@article{sintorn2008fast,

   title={Fast parallel GPU-sorting using a hybrid algorithm},

   author={Sintorn, E. and Assarsson, U.},

   journal={Journal of Parallel and Distributed Computing},

   volume={68},

   number={10},

   pages={1381–1388},

   year={2008},

   publisher={Elsevier}

}

Download Download (PDF)   View View   Source Source   

4762

views

This paper presents an algorithm for fast sorting of large lists using modern GPUs. The method achieves high speed by efficiently utilizing the parallelism of the GPU throughout the whole algorithm. Initially, GPU -based bucketsort or quicksort splits the list into enough sublists then to be sorted in parallel using merge-sort. The algorithm is of complexity n log n , and for lists of 8 M elements and using a single Geforce 8800 GTS-512, it is 2.5 times as fast as the bitonic sort algorithms, with standard complexity of n (log n ) 2 , which for a long time was considered to be the fastest for GPU sorting. It is 6 times faster than single CPU quicksort, and 10% faster than the recent GPU -based radix sort. Finally, the algorithm is further parallelized to utilize two graphics cards, resulting in yet another 1.8 times speedup.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2024 hgpu.org

All rights belong to the respective authors

Contact us: