Enabling Energy-Efficient Analysis of Massive Neural Signals Using GPGPU

• Authors: Dan Chen;Lizhe Wang;Shuaiting Wang;Muzhou Xiong;Gregor von Laszewski;Xiaoli Li
• Venue: GREENCOM-CPSCOM '10 Proceedings of the 2010 IEEE/ACM Int'l Conference on Green Computing and Communications & Int'l Conference on Cyber, Physical and Social Computing
• Year: 2010

Analysis of neural signals (such as EEG) has long been a hot topic in neuroscience community due to neural signals’ nonlinear and non-stationary features. Recent advances of experimental methods and neuroscience research have made neural signals constantly massive and analysis of these signals highly compute-intensive. Analysis of neural signals has been routinely performed upon CPU-based computer clusters with rapidly increasing scale and CPU clock speed. This inevitably incurs additional problems of excessive energy consumption, greenhouse emission, and extra cost of heat dissemination. This study proposes parallelized neural signal analysis approach based on a many-cores high performance computing technique, i.e., General-purpose computing on the graphics processing unit (GPGPU). Experimental results indicate that the GPGPU-aided approach achieved a dramatic speed-up with energy consumption minimized in contrast to using a CPU-based high-end computer cluster.