Figure 1. Performance and cost of various CPU and GPU implementations of a critical component of our model family.

Our implemented performance speed-ups for a key filtering operation in our biologically-inspired model implementation. Performance and price are shown across a collection of different GPUs, relative to a commonly used MATLAB CPU-based implementation (using a single CPU core with the filter2 function, which is coded in C++). We contrast this standard implementation with a multi-core MATLAB version, a highly-optimized C/SSE2 multi-core implementation on the same CPU, and highly-optimized GPU implementations. We have implemented speedups of over thousands of times with GPUs, resulting in qualitative changes in what kinds of model investigations are possible. More technical details and a throughout discussion of the computational framework enabling these speedups can be found in Supplemental Figure S1 and Supplemental Text S2. * These costs are based on multi-GPU systems containing four GPUs in addition to the quad-core CPU (Q9450). ** These costs include both the hardware and MATLAB yearly licenses (based on an academic discount pricing, for one year).