MapReduce-MPI Papers

If you use the MR-MPI library in one of your publications, please cite the following paper and also list the URL of the MR-MPI WWW site, namely http://www.sandia.gov/~sjplimp/mapreduce.html.

This paper describes the MR-MPI library and several MapReduce graph algorithms for informatics problems:

MapReduce in MPI for Large-Scale Graph Algorithms, S. J. Plimpton and K. D. Devine, accepted for a special issue of Parallel Computing. To appear in 2011. (abstract) (preprint)

The authors of the MR-MPI library are Steve Plimpton (sjplimp at sandia.gov) and Karen Devine (kddevin at sandia.gov).

You can send us email regarding any of these issues:

• If you use MR-MPI for something interesting, we'd like to hear about it.
• If you write a paper using the MR-MPI library, we can post a citation to it on this page.
• If you write a MapReduce program which calls the library, that others would be interested in, you can send it and we'll include it in the distribution.
• If you enhance the library itself or have ideas for doing so, we can discuss including the improvements in future releases.

This is a paper by Seung-Jin Sul and Andrey Tovchigrechko (J Craig Venter Institute) on using the MR-MPI library to parallelize two computational biology algorithms: BLAST and a neural network approach called SOM. They've run their models successfully on up to 1024 processors of the U Texas Ranger machine.

Parallelizing BLAST and SOM algorithms with MapReduce-MPI library, S.-J. Sul and A. Tovchigrechko, IEEE International Parallel & Distributed Processing HICOMB Symposium, (2011). (abstract) (download paper from HICOMB 2011 proceedings site)