Hi,
Gary, you may have noticed that I've started to port the
'multi_processor' branch to the 1.3 line (via the
'multi_processor_merge' branch). This porting is very messy and will
need to be significantly cleaned up at the end - long after all the
porting has taken place. The new design is so radically different
that the code is horribly broken at the moment. But I should be able
to fix most things.
There is one part that I can't fix though, and that is the
parallelisation of the grid search. The problem is that this code is
no longer in relax. It has all been shifted to the minfx project
(https://gna.org/projects/minfx/), a library of optimisation
algorithms. I'll skip this code for now, but it might be worth
discussing the options for having a parallelised grid search in the
1.3 line. Do you have any ideas how this could be done now that the
code spans two completely different projects?
Cheers,
Edward
Porting of the 'multi_processor' branch and the grid search parallelisation.