GPU-Based Extended Cellular Model Implementation
Alexey Emelyanov, Роман Дмитриенко · Proceedings of the Spring/Summer young researchers' colloquium on software engineering · 2010
A Cellular Automaton (Cellular Model) is a convenient and efficient way to solve a broad class of problems that belong to different areas of science, including (but not limited to) physics, math, chemistry, biology.A lot of natural phenomena and common tasks that are described via differential equations are easily modeled this way.But in spite of obvious advantages, CA-based algorithm implementations for common CPUs are highly parallel and hence extremely resource-intensive.Thus, the idea of SIMD GPU-based CA implementation looks feasible.In this paper, we will discuss the problems we faced while developing GPU-based cellular model implementation, and the ways we solved them.