CCPP Framework Meeting Minutes 2022 02 22

Agenda

• capgen unification
• Other business
• Planning for 3D Physics

---

New member

• Mike Kavulich (NCAR DTC) is joining the "big wonderful CCPP family" - welcome!

capgen unification

• One outstanding pull request for CCPP framework (capgen, #434) adding unit/system tests for automated unit conversions
  • PR review is there, Dom will work on it

Other business

• Ufuk and Dom made good progress on calling CCPP from the CMEPS mediator, related CCPP framework PR #439 is currently in draft mode
  • Ufuk can now reproduce the fluxes from the FV3 atmospheric model in the CMEPS mediator
• #419 (visualization of variable traversing through suite) is still work in progress, due in June 2022 (part of DTC's statement of work)
• Reconsider meeting time with an increasing attendance from the East Coast

Planning for 3D Physics

• What are the required semantics?
  • What operations are needed and when?
  • Exploratory - no immediate requirement, but it comes up at more and more at meetings etc (e.g. 3D PBL development at NCAR). Will also be discussed at the CCPP visioning workshop
• Two examples:
  • 3D, but vertical grid is the same everywhere - just ignoring communication between the horizontal grid columns (pseudo 3D)
  • Full 3D model: each column can have its own vertical coordinate
• Definition of vertical coordinate? Same vertical coordinate means the same as defined (and modified) by the dynamical dycore
• Requires constructing stencils (depends on what exactly is needed)
• When it is needed (horizontal communication in the time integration = run phase)?
  • If in the run phase, how can we implement this w/o impacting performance
  • Have a dedicated 3d run group or communication between the groups etc.
• We could prepare some options with pros and cons (for example in form of a white paper) to guide the discussion at the visioning workshop