Hpmoon

Hpmoon is a parallel and distributed multi-objective genetic algorithm to EEG classification. The evolutionary procedure corresponds to an island-based model whose subpopulations are distributed among the nodes of a cluster. The application is able to parallelize the evaluation of the individuals using all the CPU-GPU devices simultaneously, which provides up to 4 levels of parallelism.

Requirements

Hpmoon requires a GCC compiler and OpenCL 1.2 compliant CPU-GPU devices. It also depends on the following APIs and libraries:

• OpenMPI.
• AMD APP SDK v2.9.1 or later.
• Doxygen if you want to generate documentation.

Documentation

Hpmoon is fully documented in its Github Pages. In addition, the Makefile file contains a rule to generate Doxygen documentation in the docs/html folder.

Usage

The docs folder contains the file user_guide.pdf with the instructions necessary to use the program. You can also display help by running the program with the -h option.

Acknowledgments

This work has been funded by:

• Spanish Ministerio de Economía y Competitividad under grants number TIN2012-32039 and TIN2015-67020-P.
• Spanish Ministerio de Ciencia, Innovación y Universidades under grant number PGC2018-098813-B-C31.
• European Regional Development Fund (ERDF).

License

GNU GPLv3.

Copyright

Hpmoon © 2015 EFFICOMP.

Publications

1. J. J. Escobar, J. Ortega, A. F. Díaz, J. González, and M. Damas. "Energy-aware Load Balancing of Parallel Evolutionary Algorithms with Heavy Fitness Functions in Heterogeneous CPU-GPU Architectures". In: Concurrency and Computation: Practice and Experience 31.6 (2019), e4688. doi: 10.1002/cpe.4688
2. J. J. Escobar, J. Ortega, A. F. Díaz, J. González, and M. Damas. "Time-energy Analysis of Multi-level Parallelism in Heterogeneous Clusters: The Case of EEG Classification in BCI Tasks". In: The Journal of Supercomputing 75.7 (2019), pp. 3397-3425. doi: 10.1007/s11227-019-02908-4.
3. J. J. Escobar, J. Ortega, A. F. Díaz, J. González, and M. Damas. "A Power-Performance Perspective to Multiobjective Electroencephalogram Feature Selection on Heterogeneous Parallel Platforms". In: Journal of Computational Biology 25.8 (2018), pp. 882-893. doi: 10.1089/cmb.2018.0080.
4. J. J. Escobar, J. Ortega, J. González, M. Damas, and A. F. Díaz. "Parallel High-dimensional Multiobjective Feature Selection for EEG Classification with Dynamic Workload Balancing on CPU-GPU". In: Cluster Computing 20.3 (2017), pp. 1881-1897. doi: 10.1007/s10586-017-0980-7.