Simulation of Spiking Neural P Systems with Sparse Matrix-Vector Operations

To date, parallel simulation algorithms for spiking neural P (SNP) systems are based on a matrix representation. This way, the is implemented with linear algebra operations, which can be easily parallelized high performance computing platforms such as GPUs. Although it has been convenient first generation of GPU-based simulators, CuSNP, there some bottlenecks to sort out. For example, proposed representations SNP lead very sparse matrices, where majority values zero. It known that matrices compromise since they involve waste memory and time. problem extensively studied in literature computing. In this paper, we analyze these ideas apply them represent variants systems. We also provide new algorithm novel compressed representation matrices. conclude system variant better suits our