Um algoritmo de alto desempenho para evoluir o modelo de Potts Celular

Abstract

The simulation of cellular systems has received great interest in the last years. In particular, the cellular Potts model is widely used in the area given its precision in representing these systems. This model, in its standard form, takes a series of operations and calculations that are executed in an ine±cient way, what disables its use in large scale simulations that demand considerable time and memory for conclusion. Based on that, we propose a new algorithm of higher performance that allows to obtain results close from those obtained with the Monte Carlo algorithm in much shorter time. Techniques of concurrent execution and communication have been introduced in the algorithm, through the use of light processes for execution in computers with shared memory and using clusters of computers, respectively, aiming to reduce the processing time and making possible the execution of large scale simulations. The results presented obtained from

simulation of cellular segregation and foam evolution show a minimum s