Page tree

Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.


Mark O also mentioned that he and Yannick will be traveling to Reading UK in mid-October to visit ECMWF and to attend a workflow conference there.  The JEDI team is working on a list of questions and issues (as a Google Doc) for them to discuss with ECMWF while they are there, particularly regarding our use of eckit, fckit, ecbuild, atlas, and other tools.  Please let us know if you would like access to the Google Doc in order to add some questions/issues to bring up to ECMWF.  Mark O mentioned ecbuild in particular.  The version the version of ecbuild we're using now is behind their ECMWF's development version, both because of outstanding issues/bugs in their development version and because updating to their development version would require significant changes to all the JEDI repos and an up-to-date version of CMake.  Some specific outstanding issues with ecbuild that we plan to discuss include compatibility issues and problems with finding libraries.  

Clementine has been working with the MPI implementation of H(x) in the shallow water model using fv3-jedi as a model.


Dan has continued to work on the Poisson solver he discussed last week.  He now has it working, giving reasonably accurate results for transforming from u, v, to streamfunction/potential and back.  However, he wants to improve the accuracy by improving the interpolation. 

Dan has also been going through a substantial recent pull request by Benjamin on oops, saber, and fv3-jedi that introduces Atlas into JEDI.  Currently Benjamin is only using Atlas to provide a container for grids and fields, which will change the interface between oops and bump, and, more generally oops and saber.  There is also some discussion of implementing Atlas fields more broadly within JEDI, including possibly the increment interfaces and fv3-jedi fields.  For further details, see the feature/atlas_fields pull requests in oops, saber, and fv3-jedi, and see the discussion thread on the GitHub JEDI-Models Team.Dan also suggested that we add atlas to the jedi stack so we don't have to install it together with the bundle.  He was finding that the compilation and ctests take a significant amount of time.