High Performance and Increased Precision Techniques for Feynman Loop Integrals