<div dir="ltr"><div>Exactly, I personally look for:<br><br>1) type checking during the
function calls. This helps also not looking into the manual for every single petsc function call. <br><br>2) low level access to petsc data structure. For example in my own code I already have my own hadnling of parallel vectors and matrices. When I use petsc, I have to always copy in/out all the data which is not optimal. So, it would be great if I can share or be change the petsc data in my Fortran code i.e. not through the petsc function calls. I understand this might be dangerous or not duable etc. But, I see no other solution. <br>
<br>3) easy
definitions of petsc namespace through the Fortran modules, so I can
choose what to import (using the Fortran 'only' statement). <br><br></div>There
is no general automatic interface generation for Fortran, as much as
I'm aware of. The gtk-fortran project, they wrote a python script to
generate this interface automatically and it really works nicely. I
could not imagine writiing such nice GTK GUI with Fortran. <br><br><br></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Mon, Oct 28, 2013 at 2:15 PM, Jed Brown <span dir="ltr"><<a href="mailto:jedbrown@mcs.anl.gov" target="_blank">jedbrown@mcs.anl.gov</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">Hossein Talebi <<a href="mailto:talebi.hossein@gmail.com">talebi.hossein@gmail.com</a>> writes:<br>
> What was in my mind is to create an interface additional to what is already<br>
> there in petsc. So this should be compiled along with anyone's code if the<br>
> used ocmpiler supports F03 or the necessary features of F03.<br>
<br>
</div>I don't think anyone will want to manually maintain the interface<br>
definitions for all of PETSc. (F2003 users can call the C interface<br>
using the C types support now, but of course you'd rather have Fortran<br>
interfaces so that you can have type checking.) It seems to me that the<br>
big win of F2003 is that it should be possible to automatically generate<br>
almost all these definitions. But how would you manage callbacks?<br>
</blockquote></div><br><br clear="all"><br>-- <br><div dir="ltr"><div>Dr.-Ing. Hossein Talebi</div><div><br></div><div>Bauhaus-Universität Weimar</div><div>Graudate Research School 1462 "Evaluation of Coupled Numerical and Experimental Partial Models in Structural Engineering " </div>
<div>Postdoctoral Researcher</div><div><br></div><div>Berkaer Str. 9</div><div>99425 Weimar</div><div>Tel.: +49/3643/58 41 03</div><div>Fax: +49/3643/58 41 01</div><div>E-Mail: hossein.talebi<a href="mailto:maximilian.huber@uni-weimar.de" target="_blank">@uni-weimar.de</a></div>
<div><a href="http://www.uni-weimar.de/grk1462" target="_blank">www.uni-weimar.de/grk1462</a></div></div>
</div>