Multi-GPU MapReduce on GPU Clusters
Authors:
- Jeff A. StuartJSJeff A. Stuart
- John D. OwensJOJohn D. Owens
- 1 May 2011
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15302075,p. 1068-1079
Abstract
We present GPMR, our stand-alone MapReduce library that leverages the power of GPU clusters for large-scale computing. To better utilize the GPU, we modify MapReduce by combining large amounts of map and reduce items into chunks and using partial reductions and accumulation. We use persistent map and reduce tasks and stress aspects of GPMR with a set of standard MapReduce benchmarks. We run these benchmarks on a GPU cluster and achieve desirable speedup and efficiency for all benchmarks. We compare our implementation to the current-best GPU-MapReduce library (runs only on a solo GPU) and a highly-optimized multi-core MapReduce to show the power of GPMR. We demonstrate how typical MapReduce tasks are easily modified to fit into GPMR and leverage a GPU cluster. We highlight how total and relative amounts of communication affect GPMR. We conclude with an exposition on the types of MapReduce tasks well-suited to GPMR, and why some tasks need more modifications than others to work well with GPMR.Keywords
This publication has 15 references indexed in Scilit:
- MapCG2010
- Understanding the efficiency of ray traversal on GPUs2009
- Sector and Sphere: the design and implementation of a high-performance data cloudPhilosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2009
- CellMR: A framework for supporting mapreduce on asymmetric cell-based clusters2009
- Map-reduce as a Programming Model for Custom Computing Machines2008
- Scalable Parallel Programming with CUDAQueue, 2008
- Google’s MapReduce programming model — RevisitedScience of Computer Programming, 2008
- MapReduceCommunications of the ACM, 2008
- Map-reduce-merge2007
- Minimal perfect hash functions made simpleCommunications of the ACM, 1980