Strategies to parallelize a �nite element mesh truncation technique on multi- and manycore architectures