Hi Danilo,
ok, sorry just got what happened… rad -N will render all views
in parallel (one process per view), if multiple views are defined. In
your case, that is six rendering processes, one for each view, plus one
control process. If only one view is defined, it will split the view
into tiles (internally by calling rpiece) and render the tiles in parallel.
So if you want to use all cores, and if you have more cores than views
to render, you should render only one view at a time calling
rad -N -v
This will render only one view, but split the task (internally it calls
rpiece) so that all cores will share the load. So you would call rad
sequentially, e.g. by looping over a list of views. In bash (untested):
for view in “top” “left” “right” “front” “back”; do rad -N 16 -v $view
render.rif; done
This assumes that the given, named views are defined in a control-file
render.rif. And obviously all this makes sense only if it takes a
significant time to complete one view…
Best, Lars.