[jd210]

Hi All,

I'm trying to write a new smoother, mainly for use with GAMG, after writing it I discovered it wasn't performing how I expected. I narrowed the problem down to it actually being an issue with it being a new smoother.

To try to explain it better I have copied the DIC smoother meaning I have the original openfoam called "DIC" then I have "myDIC" which is loaded at runtime in the same way as with a new linear solver. However despite the code being exactly the same, only the names have changed, they perform completely differently. Using the exact same case so the only thing that is changed is using DIC or myDIC to solve for p_rgh takes much longer.

With the original OpenFOAM DIC solving for p_rgh takes 2, 2 and 5 iterations in the first time step. Using myDIC 4, 5 and 16 iterations are required with the final residuals being larger.

Can anyone think of a reason why this may happen? It doesn't make sense to me.

Thanks

Josh