8780

Acceleration of Selective Cationic Antibacterial Peptides computation: A comparison of FPGA and GPU approaches

D. Garcia-Ordaz, M. Arias-Estrada, M. Nuno-Maganda, C. Polanco, G. Del Rio
National Institute for Astrophysics, Optics and Elecronics, Computer Science Dept. Apdo. Postal 51 and 216. Puebla, Pue. 72000 Mexico
International Supercomputing Conference, 2012

@article{garcia2012acceleration,

   title={Acceleration of Selective Cationic Antibacterial Peptides computation: A comparison of FPGA and GPU approaches},

   author={Garcia-Ordaz, D. and Arias-Estrada, M. and Nu{~n}o-Maganda, M. and Polanco, C. and Del Rio, G.},

   year={2012}

}

Download Download (PDF)   View View   Source Source   

1904

views

Prediction of physicochemical properties of peptide sequences can be used for the identification of "Selective Cationic Amphipatic Antibacterial Peptides" (SCAAP), with possible applications in different diseases treatment. The exhaustive computation of physicochemical properties of peptide sequences can lead to reduce the search space of SCAAP, but the combinatorial complexity of these calculations is a high-performance computing problem. There are several alternatives to deal with the high computational demand to accelerate computation, in particular the use of supercomputer resources has been the traditional approach but other computational alternatives are cost-effective to bring the supercomputing power to the desktop. In this work we compare acceleration and performance of SCAAP (with 9 amino acids length), computation among three different platforms: traditional PC computation, FPGA acceleration by custom architecture implementation, and GPU acceleration with CUDA-C programming. The comparison is carried out with four physicochemical properties codes used to identify peptide sequences with potential selective antibacterial activity. The FPGA acceleration reaches 100x speedup compared to a PC computer and the CUDA-C implementation shows a performance 1000x speedup at desktop level with the potential of increasing the number of peptides being explored. A discussion of performance and tradeoffs for the implementation on each platform is given.
No votes yet.
Please wait...

* * *

* * *

HGPU group © 2010-2024 hgpu.org

All rights belong to the respective authors

Contact us: