Extending a Highly Parallel Data Mining