A parallel version of the cellular automata static recrystallization model dedicated for high performance computing platforms - Development and verification

Development and verification of a parallel version of the micro-scale cellular automata (CA) static re-crystallization (SRX) model is the main goal of the present research. Firstly, a framework of sequential CA-SRX model is presented. Then, crucial modifications are developed and introduced to the model to handle parallel execution. A particular attention is focused on the development of computational domain decomposition schemes and information gathering modules from the CA subdomains. The computational capabilities of the developed model are evaluated with the classical parallelization indicators, i.e., computational speedup and efficiency. Finally, the results from a parallel version of the CA-SRX code are validated against results from the sequential approach, to prove that the new model does not introduce any unphysical artefact.