Closure-specific temporary variables.

[ali-ramadhan]

As discussed with @jm-c, I think another feature we will need is the ability to design closure-specific temporary variables.

For example, the implementation of Constant Smagorinsky or Anisotropic Minimum Dissipation will benefit from the ability to compute the eddy viscosity (at 4 locations in a cell --- 3 FaceFields and 1 CellField) and re-use it in calculating the contribution of the subgrid closure to the momentum 'source terms'.

Originally posted by @glwagner in https://github.com/ali-ramadhan/Oceananigans.jl/issues/73#issuecomment-467660286

[ali-ramadhan]

I think we might be able to get away with recomputing, especially on the GPU, so closing this issue. Also not an actionable issue.

[johncmarshall54]

Putting in div.v/delta t in the pressure solve is an urgent priority, as it a test for the poisson solve.

…

On Wed, May 22, 2019, 7:24 AM Ali Ramadhan ***@***.***> wrote: I think we might be able to get away with recomputing so closing this issue. Also not an actionable issue. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#94>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AKXUEQQJKFBRSAHJ35F5DTLPWUUODANCNFSM4G24CTZQ> .

[ali-ramadhan]

@johncmarshall54 Agreed. Working on it as part of issue #161.