Remove MPI bottlenecks in grid initialization
This is a WIP merge request, based on the work of Rogeli Grima (@rgrima) from BSC, and (almost, see below) rebased on master
.
The code uses a new asynchronous communication scheme.
This code was not merged in the MaX branches, as originally intended, because of a minor missing functionality, and now it needs to be updated to fix conflicts with the 'nsm' update to master.
ToDo:
-
Fix conflicts with master (see commit 1aea8c50) -
Make sure that the new scheme is applied to all relevant routines (in particular, to MatrixMtoOC) -
Clarify changes in m_timer.F90 ("ORIG" way of reporting vs new one) -
Testing
Edited by Alberto Garcia