Integrating AD with Object-Oriented Toolkits for High-Performance Scientific Computing

摘要

Often the most robust and efficient algorithms for the solution of large-scale problems involving nonlinear PDEs and optimization require the computation of derivatives. We examine the use of automatic differentiation (AD) for computing first and second derivatives in conjunction with two parallel toolkits, the Portable, Extensible Toolkit for Scientific Computing (PETSc) and the Toolkit for Advanced Optimization (TAO). We discuss how the use of mathematical abstractions in PETSc and TAO facilitates the use of AD to automatically generate derivative codes and present performance data demonstrating the suitability of this approach.