Hi,<br><br>for this approach, the best would be to<br>(before the multi-process job)<br>- put your map into the PERMANENT mapset<br><br>(for each process in parallel)<br>- make a new mapset for each process<br>- define the region of each mapset as a chunk of the whole region<br>
- make your calculation<br><br>(once the process competed)<br>- put together all the results (eg, with r.patch)<br><br><br>Sylvain<br><br><br>