Efficient Collective Communication Paradigms for Hyperspectral Imaging Algorithms Using HeteroMPI David Valencia 1,⋆ , Antonio Plaza 1 , Vladimir Rychkov 2 , and Alexey Lastovetsky 2 1 Technology of Computers and Communications Dept.,Technical School of C´aceres University of Extremadura, Avda. de la Universidad S/N, E-10071 C´aceres (Spain) {davaleco,aplaza}@unex.es 2 Heterogeneous Computing Laboratory, School of Computer Science and Informatics University College Dublin, Belfield, Dublin 4, Ireland {vrychkov,alastovetsky}@ucd.ie Abstract. Most of the parallel strategies used for information extrac- tion in remotely sensed hyperspectral imaging applications have been im- plemented in the form of parallel algorithms on both homogeneous and heterogeneous networks of computers. In this paper, we develop a study on efficient collective communications based on the usage of HeteroMPI for a parallel heterogeneous hyperspectral imaging algorithm which uses concepts of mathematical morphology. Keywords: Hyperspectral Imaging Algorithms, HeteroMPI. 1 Introduction Hyperspectral imaging identifies materials and objects in the air, land and water on the basis of the unique reflectance patterns that result from the interaction of solar energy with the molecular structure of the material[1]. Most applications of this technology require timely responses for swift decisions which depend upon high computing performance of algorithm analysis. Examples include target de- tection for military and defense/security deployment, urban planning and man- agement, risk/hazard prevention and response including wild-land fire tracking, biological threat detection, monitoring of oil spills and other types of chemical contamination. These images are characterized by covering tens or even hundreds of kilometers long, having hundreds of MB in size. Few consolidated parallel tech- niques for analyzing this kind of data currently exist in the open literature, and mainly all of them implemented on homogeneous networks of computers using MPI. Although the standard MPI[3] has been widely used to implement paral- lel algorithms for Heterogeneous Networks of Computers (HNOCs), it does not provide specific means to address some additional challenges posed by these net- works, including the distribution of computations and communications unevenly, ⋆ Corresponding author. A. Lastovetsky et al. (Eds.): EuroPVM/MPI 2008, LNCS 5205, pp. 326–331, 2008. c  Springer-Verlag Berlin Heidelberg 2008