Energy-aware acceleration on GPUs: Findings on a bioinformatics benchmark

Neuroimage
Authors

Juan Pérez

Andrés Rodríguez

Juan Francisco Chico

Domingo López-Rodríguez

Manuel Ujaldón

Published

1 December 2018

Publication details

Sustainable Computing: Informatics and Systems vol. 20, pp. 88 – 101

Links

DOI

 

Abstract

This paper performs a complete study on performance and energy efficiency of biomedical codes when accelerated on GPUs (Graphics Processing Units). We have selected a benchmark composed of three different building blocks which constitute the pillars of four popular biomedical applications: Q-norm, for the quantile normalization of gene expressions, reg f3d, for the registration of 3D images within the NiftyReg library, bedpostx (from the FSL neuroimaging package) and a multi-tensor tractography for the analysis of diffusion images. We try to identify (1) potential scenarios where performance per watt can be optimal in large-scale biomedical applications, and (2) the ideal GPU platform among a wide range of models, including low power Tegras, popular GeForces and high-end Titans. Experimental results conclude that data locality and arithmetic intensity represent the most rewarding ways on the road to high performance bioinformatics when power is a major concern.

Cites

The following graph plots the number of cites received by this work from its publication, on a yearly basis.

Citation

Please, cite this work as:

[Pér+18] J. Pérez, A. Rodríguez, J. F. Chico, et al. “Energy-aware acceleration on GPUs: findings on a bioinformatics benchmark”. In: Sustainable Computing: Informatics and Systems 20 (2018), pp. 88-101. DOI: 10.1016/j.suscom.2018.01.001.

@article{perez2018energy,
     title={Energy-aware acceleration on GPUs: findings on a bioinformatics benchmark},
     author={Pérez, Jesús and Rodríguez, Andrés and Chico, Juan Francisco and López-Rodríguez, Domingo and Ujaldón, Manuel},
     journal={Sustainable Computing: Informatics and Systems},
     volume={20},
     pages={88–101},
     year={2018},
     publisher={Elsevier},
     doi={10.1016/j.suscom.2018.01.001}
}