Publication details

<< back
 

Computing Strongly Connected Components in Parallel on CUDA

Basic information
Original title:Computing Strongly Connected Components in Parallel on CUDA
Authors:Jiří Barnat, Petr Bauch, Luboš Brim, Milan Češka
Further information
Citation:BARNAT, Jiří - BAUCH, Petr - BRIM, Luboš - ČEŠKA, Milan. Computing Strongly Connected Components in Parallel on CUDA. In Proceedings of 25th IEEE International Parallel & Distributed Processing Symposium. Anchorage, AK : IEEE Computer Society, 2011. ISBN 978-1-61284-372-8, pp. 544 - 555.
Original language:English
Field:Informatika
WWW:link to a new windowhttp://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6012868
Type:Article in Proceedings
Keywords:parallel graph algorithms; strongly connected components; CUDA

The problem of decomposing a directed graph into its strongly connected components is a fundamental graph problem inherently present in many scientific and commercial applications. In this paper we show how some of the existing parallel algorithms can be reformulated in order to be accelerated by NVIDIA CUDA technology. In particular, we design a new CUDA-aware procedure for pivot selection and we adapt the particular parallel algorithms for CUDA accelerated computation. We also experimentally demonstrate that with a single GTX 480 GPU card we can easily outperform optimal serial CPU implementation -- by an order of magnitude in most cases, 40 times on some sufficiently big instances. This is a particularly interesting result as unlike the serial CPU case, the asymptotic complexity of the parallel algorithms is not optimal.

Related projects:

 
 
top