Running ROMS 2.2 in 3D with MPI and partitioning in the I-direction, I get significant discontinuities in the vertical velocities at the tile boundaries.
The problem disappears if the model is tiled only in the J-direction.
Has anyone else noticed this?
Deborah
Problem with vertical velocities at tile boundaries
-
debcox
- Posts: 15
- Joined: Tue Mar 09, 2004 3:21 pm
- Location: University of New South Wales, Australia
Re: Problem with vertical velocities at tile boundaries
After a bit more checking, I think my problem was caused by some side sponge layers I had inadvertently left in from a previous (single tile) run. When I had more than one tile in the I-direction they were being applied at the sides of every tile. Sorry about that! 
I guess it shows there is a need to be careful how things are defined when using MPI ...
I guess it shows there is a need to be careful how things are defined when using MPI ...