Skip to content

CCPP Framework Meeting Minutes 2022 11 15

Michael Kavulich edited this page Nov 15, 2022 · 7 revisions

Agenda


Attendees: Mike Kavulich, Cheryl, Ligia, Lulin, Dom, Jordan, Grant, Dustin, Courtney

GitHub issues

  • None new.

Update from Cheryl on CCPP in CGD

  • CGD has officially announced that CAM7 will adopt CCPP to manage their physics
  • Likely starting next year. Officially "January" but there is no CCPP framework person in CGD.
    • Steve will be offered contract work, but will not be full time
    • Starting the process of opening a new position
  • Let's have more discussion to work together on getting CGD and DTC interests converged!

Interstitial Scheme Changes

There has been A LOT of discussion about improving the interoperability of interstitial schemes amongst both DTC and the larger CCPP Physics Management group. In addition, there has been a longstanding desire to reduce the amount of interstitial schemes needed to make a working suite. Therefore:

  • DTC team has been examining the current crop of interstitial schemes with a critical eye toward reducing/reorganizing them: See https://docs.google.com/spreadsheets/d/1_Zw5B49QM_HHgMhstwz_8xFkG9ygxvAkxKmpm2ECP28/edit#gid=368751223
  • Some of the code found in interstitial schemes should be able to be handled by autogenerated code from the framework
  • Key functionalities to move to the framework include:
    • Tendency handling and application; all schemes should return tendencies; SDFs can specify process-/time-split and the framework "does the right thing"
    • Constructing arrays of constituents to be operated on by schemes (e.g. vertically-diffused tracers, convectively-transported tracers, ...)
    • Automatically handling array size expansion (e.g. host provides (i,k) arrays, but scheme needs (i,j,k) arrays)
    • Taking over what is in GFS_debug.F90 (writing out all vars used by current suite -- framework knows the location in memory of all variables, so why not?)

Discussion notes

CGD has a "wishlist" to get "nested suite" capability incorporated

  • This would help simplify/shorten suite definition files

CGD also wants the process/time split capability!

Cheryl is discussing the possibility of compiling a list of "CAM" variable names and CCPP standard names in the lead up to this CAM--> CCPP conversion

  • CAM has a more complicated set of variables compared to UFS
    • State variables
    • Pbuf (essentially a "common block" of variables)
    • Will work on mapping the names of these existing names to the equivalent CCPP standard names
    • Grant and Dom have a starting point for this based on past CESM/CAM work (for one scheme at least)

Other business

  • Mike began updating ccpp-framework Wiki pages, still a work in progress
  • Mike is experimenting with adding more pytest tests to tests/ directory, will document the process for future users.

Notes

Clone this wiki locally