open TELEMAC-MASCARET The mathematically superior suite of solvers

I've recently been running a simulation using the dragfo subroutine in scalar. This is fine for a small run. However, because of the size of my model (~140,000 nodes) I need to run the model in parallel. However, dragfo does not seem to run in parallel. Running in parallel on one processor is fine, but increasing it affects the partitioning of the mesh.

Does anyone have a suggestion of how to ensure, when the mesh is partitioned, it does not split the area defined in dragfo for the command: INPOLY(X(I),Y(I),XSOM,YSOM,NSOM).

Regards

David

Hello,

Officially dragfo should work in parallel, since the area is summed over all processors with function P_DSUM, and INPOLY should work as well because all coordinates are unchanged by partitioning, moreover we use it here in parallel for wind turbines, so it could be something else in your case, unless we were lucky and nobody really splits the zone with drag forces...

Regards,

JMH

Thanks Jean-Michel, I will dig a little a deeper and if I cant find a solution I will post again.

Kind Regards

David

Thank you Jean-Michel. After investigating further I found that the example dragfo in doxydocs has been updated since I first used the subroutine and now includes the line:

IF(NCSIZE.GT.0) AIRE=P_DSUM(AIRE)

My code works now.

Thank you again.