Moorkanikara J, Felch A, Chandrashekar A, Dutt N, Granger R, Nicolau A, Veidenbaum A (2009) Brain derived vision algorithm on high performance architectures. Int Journ Parallel Prog 37:345-369.

Moorkanikara J, Felch A, Chandrashekar A, Dutt N, Granger R, Nicolau A, Veidenbaum A

Abstract

Even though computing systems have increased the number of transis-tors, the switching speed, and the number of processors, most programs exhibit limitedspeedup due to the serial dependencies of existing algorithms. Analysis of intrinsically parallel systems such as brain circuitry have led to the identification of novel architec-ture designs, and also new algorithms than can exploit the features of modern multipro-cessor systems. In this article we describe the details of a brain derived vision (BDV)algorithm that is derived from the anatomical structure, and physiological operatingprinciples of thalamo-cortical brain circuits. We show that many characteristics of theBDV algorithm lend themselves to implementation on IBM CELL architecture, andyield impressive speedups that equal or exceed the performance of specialized solu-tions such as FPGAs. Mapping this algorithm to the IBM CELL is non-trivial, and wesuggest various approaches to deal with parallelism, task granularity, communication,and memory locality. We also show that a cluster of three PS3s (or more) containingIBM CELL processors provides a promising platform for brain derived algorithms,exhibiting speedup of more than 140×over a desktop PC implementation, and thusenabling real-time object recognition for robotic systems.