Transformation of Algorithms in Fortran (TAF) Demonstrator

TAF is a source-to-source translator for modern Fortran, which takes Fortran source code as input, applies one of several semantic transformations, and generates according Fortran output. TAF supports several semantic transformations, the most important of which is Automatic Differentiation (AD) that generates code to compute the first derivative (Jacobian matrix). The generated derivative code can operate in forward (tangent linear) or reverse (adjoint) mode. Higher-order derivative code can be generated by applying TAF to previously generated lower-order derivative code. The form below provides a simplified and limited interface for demonstrating and exploring the capabilities of TAF.

Your Fortran code

Your TAF-generated code

TAF options

Differentiation

 
 

Language

   

Copyright © FastOpt - all rights reserved