Exploring data and thread parallelism to achieve energy efficiency from the software developer's perspective

• Rebeka G. de Oliveira UFPE
• Sérgio V. Cavalcante UFPE

Abstract

Environmental issues and the demand for mobility in mobile devices has generated a growing demand for reduction in energy consumption. Parallelism is a bet on energy efficiency in computing devices. This study aims to examine the use of thread parallelism and data parallelism for the purpose of improving energy efficiency from the point of view of the software developer. A case study was conducted with the benchmark ALPBench, which was analyzed the performance and power consumption of applications. The reduction in energy consumption reached 77.1% and the gain in performance reached 75.5%. The results provide further evidence that through parallelism is possible to achieve the energy efficiency of software applications.

References

Salapura V., Bickford R., Blumrich M., Bright A., Chen D., Coteus P., Gara A., Giampapa M., Gschwind M., Gupta M., Hall S., Haring R., Heidelberger P., Hoenicke D., Kopcsay G., Ohmacht M., Rand R., Takken T., Vranas P.; Power and performance optimization at the system level, 2nd conference on Computing frontiers. Italy, 2005.

Li, M., Sasanka, R., Adve, S., Chen, Y., Debes, E. “The ALPBench benchmark suite for multimedia applications”. In IEEE Intl. Symp. on Workload Characterization, 2005.

Tanenbaum, Andrew S., Modern operating systems – 2nd ed. Prentice-Hall, Inc., Upper Saddle River, New Jersey, USA, 2001.

Aki, Selim G. The design and analysis of parallel algorithms. Prentice-Hall, Inc. Upper Saddle River, New Jersey, USA, 1989.

Intel Corporation. The IA-32 Intel Architecture Optimization Reference Manual, 2004

Šimunić, T.; Benini, L.; Micheli G.; Hans M. “Source code optimization and profiling of energy consumption in embedded systems”, Proceedings of the 13th international symposium on System synthesis, Setembro, 2000, Madrid, Espanha.

Sasanka, R., Li, M., Adve, S., Chen, Y., Debes E. “ALP: Efficient support for all levels of parallelism for complex media applications,” ACM Transactions on Architecture and Code Optimization, Vol. 4, No. 1, Art. 3. Março, 2007.

Markus Lorenz, Peter Marwedel, Thorsten Drager, Gehard Fettweis, Rainer Leupers. “Compiler based Exploration of DSP Energy Savings by SIMD Operations”, XXX, 2004.

Júnior, Meuse Nogueira de Oliveira. “Estimativa do Consumo de Energia Devido ao Software: Uma abordagem baseada em redes de Petri coloridas”. 214p. Tese de 1788 Doutorado, Pós-Graduação em Ciência da Computação, Universidade Federal de Pernambuco. Recife, Outubro de 2006.