Skip to content

Home

Jump to bottom
Guillaume Roullet edited this page Dec 1, 2019 · 4 revisions

Welcome to the Nyles wiki!

Now that we have a stable non-linear model (11/30/19), here are the things we want to implement

  1. periodic channel This requires to handle the halo. We should do that with still a single domain (no subdomains). Once this is done we can start doing jet stability (shear, centrifugal, symmetric, Kelvin-Helmholz).

  2. integral diagnostics We want to monitor kinetic and potential energy, and also tracer mean and standard deviation

  3. pass grid to MG (DONE) To drop the assumption that dx=dy=dz=1.

  4. vorticity and BC Carefully handle the BC (i=-1, j=-1) in fortran_vorticity.f90. So far it's set to 0, which is wrong.

  5. add forcing on tracer We want to do Raleigh Benard convection in 3D.

  6. full domain decomposition Once the channel works this shouldn't be long to do. The main piece is to implement the glue/split in MG. We can test Nyles on bigger grids, O(32**3) is cool but we definitely want to have results at larger resolution.

Clone this wiki locally