[ADOL-C] using adolc with openmp

[Andrea Walther]

Hi Ian,


> Possibly the tapeless forward AD tools of ADOLC are better suited to  
> my application;

yes, this seems to fit better to your application.

> however, I need to generate sparse Jacobian matrices and it is  
> unclear from documentation how to do this. It seems only dense  
> derivatives via the tapeless forward mode are possible?


So far there is no automated provision for a sparse tapeless forward mode.
However, this could be arrange manually in the following way:

If one knows the sparsity structure of the Jacobians von can generate
a so-called seed matrix S. This could for example be done in a preprocessing
step using the tape-based version. Having this seed matrix at hand,
one can then use the tapeless vector forward mode of ADOL-C
to compute the Jacobian matrix product

J * S

which should much more efficient than computing the Jacobian in a dense
fashion. Subsequently one can then reconstruct the Jacobian from its
compressed form J*S

This is done automatically in the sparse Drivers of the tapebased ADOL-C
but also illustrated in one of the examples on sparsity.

If you need help to rebuild this process for the tapeless version we  
would be happy to support you.

Best regards

Andrea

-- 
Prof. Dr. Andrea Walther
Lehrstuhl fuer Mathematik und ihre Anwendungen
Institut fuer Mathematik
Universitaet Paderborn
Warburger Str. 100
33098 Paderborn

Email: andrea.walther at uni-paderborn.de
Phone: ++49 5251 602721
        ++49 5251 602724 (sekr.)
Fax:   ++49 5251 603728

**********