Optimizing the Linear Fascicle Evaluation Algorithm for Multi-Core and Many-Core Systems