Performance and energy efficiency evaluation of heterogeneous systems for bioinformatics

• Enzo Rucci ^[1]
  1. [1] Universidad Nacional de La Plata

     Universidad Nacional de La Plata

     Argentina

     
• Localización: Journal of Computer Science and Technology, ISSN-e 1666-6038, Vol. 16, Nº. 2, 2016, págs. 104-105
• Idioma: inglés
• Enlaces
  • Texto completo (pdf)
• Resumen

  • The power consumption problem represents one of the major obstacles for exascale systems design. As a consequence, the scientific community is searching for different ways to improve power efficiency of HighPerformance Computing (HPC) systems [1]. One recent trend to increase compute power and, at the same time, limit power consumption of these systems lies in adding accelerators and co-processors, like NVIDIA/AMD graphic processing units (GPU) or Intel Xeon Phi co-processors. On the other hand, field programmable gate arrays (FPGA) stands as a promissory alternative due to their increasing computational capability, low power consumption and the development of new tools that reduce programming cost. These hybrid systems that use different processing resources are called heterogeneous systems and area capable of achieving better FLOPS/Watt ratios [2]. Bioinformatics is one of the areas affected by current HPC problems due to the exponential growth of biological data in the last years and the increasing number of bioinformatics applications demanding HPC to meet performance requirements. One of these applications is sequence alignment, which is considered to be fundamental procedure in biological sciences [3]. The alignment process compares two or more biological sequences and its purpose is to identify regions of similarity among them. The Smith-Waterman (SW) algorithm [4] is a popular method for local sequence alignment that has been used as the basis for many subsequent algorithms, and is often employed as a benchmark when comparing different alignment techniques. However, due to the quadratic computational complexity of Smith-Waterman algorithm, several heuristics are used in practice that reduce the execution time but at the expense of not guaranteeing to discover the optimal local alignments.