Clúster de Computación Científica de Bajo Coste y Consumo

  1. M. Hernández
  2. A.A. Del Barrio
  3. G. Botella
Journal:
Enseñanza y aprendizaje de ingeniería de computadores: Revista de Experiencias Docentes en Ingeniería de Computadores

ISSN: 2173-8688

Year of publication: 2018

Issue: 8

Pages: 85-96

Type: Article

DIALNET GOOGLE SCHOLAR HANDLE: https://hdl.handle.net/10481/53301

More publications in: Enseñanza y aprendizaje de ingeniería de computadores: Revista de Experiencias Docentes en Ingeniería de Computadores

Sustainable development goals

info

SDG classification obtained using Aurora SDG artificial intelligence model.

Abstract

In this work we present the construction of a cluster based on low-end FPGAs. Such cluster is able to execute data-intensive applications in a similar amount of time as a 56-threaded workstation, which is much more power-hungry and expensive. In order to perform this task, an in-house Debian 8-based image has been developed, and the required software to run OpenCL kernels compiled with Intel FPGA SDK for OpenCLv16.0 has been installed. Moreover, several comparisons tackling execution time, cost and energy consumption have been performed. As a result, we concluded that the cluster is almost 6 times cheaper than the workstation, and reaching 83% energy reductions.

Bibliographic References

  • Michael Feldman 2017. “Microsoft’s Plan for FPGAs in Azure”. https://www.top500.org/news/ microsofts-plans-for-fpgas-in-azure-should-worrytraditional-chipmakers/. [Online; accedido 09-Abril-2018].
  • George Leopold 2016. “Intel’s FPGAs Target Datacenters, Networking”. https://www.hpcwire.com/2016/ 10/06/intels-fpgas-target-datacenters-networking/. [Online; accedido 09-Abril-2018].
  • Máster en Ingeniería Informática de la Universidad Complutense de Madrid, http://informatica.ucm.es/estudios/2017-18/master-ingenieriainformatica
  • D. Lora et al., “Sistema de Seguridad Basado en una Plataforma Heterogénea Distribuida“, Enseñanza y Aprendizaje de Ingeniería de Computadores, 5: 29-38 (2015).
  • F. Parrales et al. “Una Orquesta Sinfónica como Ejemplo de Aplicación de un Sistema Empotrado Distribuido”, Enseñanza y Aprendizaje de Ingeniería de Computadores, 5: 115- 124 (2015).
  • I.M. Laclaustra et al. “Sistema Domótico Distribuido para Controlar el Riego y el Aire Acondicionado en el Hogar”, Enseñanza y Aprendizaje de Ingeniería de Computadores, 6: 87-102 (2016).
  • H. Ivanov et al. “Bomberman modo multijugador”, Enseñanza y Aprendizaje de Ingeniería de Computadores, 7: 53-68 (2017).
  • Jesús Martín Alonso et al., 2016. A distributed HW-SW platform for fireworks. Proceedings of the Summer Computer Simulation Conference (SCSC '16). Society for Computer Simulation International, Montreal, artículo 17, 7 pages.
  • A. A. Del Barrio et al., "A Distributed Clustered Architecture to Tackle Delay Variations in Datapath Synthesis," IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 35, no. 3, pp. 419-432, 2016.
  • Bai, X. et al. 2017. “Acceleration of RSA processes based on hybrid ARM-FPGA cluster”. IEEE Symposium on Computers and Communications (ISCC), pp. 682–688.
  • Castillo, J. et al. 2009. “Hardware accelerated montecarlo financial simulation over low cost FPGA cluster”. IEEE International Symposium on Parallel Distributed Processing, pp. 1–8.
  • Domingo, R. et al. 2017. “High-level design using Intel FPGA OpenCL: A hyperspectral imaging spatialspectral classifier”. International Symposium on Reconfigurable Communication-centric Systems-on-Chip (ReCoSoC), pp. 1–8.
  • A. A. Del Barrio and R. Hermida, "A slack-based approach to efficiently deploy radix 8 booth multipliers," Design, Automation & Test in Europe Conference & Exhibition (DATE), 2017, Lausanne, 2017, pp. 1153-1158.
  • Fernández, D. et al. 2018. “Complexity reduction in the HEVC/H265 standard based on smooth region classification”. Digital Signal Processing vol. 73, pp. 24 – 39.
  • Fernández, D. G. et al. 2018. “Fast and effective CU size decision based on spatial and temporal homogeneity detection”. Multimedia Tools and Applications vol. 77 (5), pp. 5907– 5927.
  • A. A. D. Barrio et al. 2016, "A Partial Carry-Save On-the-Fly Correction Multispeculative Multiplier," IEEE Transactions on Computers, vol. 65, no. 11, pp. 3251-3264.
  • Hernández, M. 2017. “Creación de un clúster de computación científica basado en FPGAs de bajo coste y consumo”. Master’s thesis, Universidad Complutense de Madrid, Spain. http://eprints.ucm.es/43993/.
  • Mariano Hernández 2017. “Debian-8 with OpenCL for DE1-SoC”. https://github.com/mariano2AA3/ tfm-de1soc-opencl-cluster. [Online; accedido 20-Marzo2018].
  • Khronos Group 2018. “OpenCL: The Open Standard for Parallel Programming of Heterogeneous Systems”. www.khronos.org/opencl. [Online; accedido 06-Febrero-2018].
  • Mencer, O. et al. 2009. “Cube: A 512-FPGA cluster”. Southern Conference on Programmable Logic (SPL), pp. 51–57.
  • Meyer-Baese, U. 2014. Digital Signal Processing with Field Programmable Gate Arrays. 4th ed. Springer.
  • Sass, R. et al. 2007. “Reconfigurable Computing Cluster (RCC) Project: Investigating the Feasibility of FPGA-Based Petascale Computing”. IEEE Symposium on FieldProgrammable Custom Computing Machines (FCCM 2007), pp. 127–140.
  • Kim, M. S. et al. 2018. “Low-power implementation of Mitchell’s approximate logarithmic multiplication for convolutional neural networks”. Asia and South Pacific Design Automation Conference (ASP-DAC), pp. 617–622.
  • Outervision. “Power Calculator”. https://outervision.com/power-supply-calculator. [Online; accedido 20-Marzo-2017].
Data source: Dialnet