Skip to content
This repository has been archived by the owner on Oct 23, 2020. It is now read-only.

Agenda Notes Monday March 13th 2017 11:00 AM

Jump to bottom
Michael Duda edited this page Mar 13, 2017 · 9 revisions

MPAS-Developer Telecon:

Date: 13 March 2017
Time: 11:00 MDT / 19:00 CET

Call-in number: 1-866-740-1260
Access Code: 4978161

Agenda Topics:

  1. Framework modifications:
  2. Scoping and brainstorming: How can we envision a subset I/O capability looking from a user perspective?
    • MPAS-Atmosphere will soon need an efficient way to read grid cells/edges/vertices only along the perimeter of a large region
    • Would this capability be useful to output, e.g., a single ocean basin?
    • Handling I/O on a subset of mesh elements may not be too difficult in the mpas_io module
    • How might users and developers describe the subset of elements
  3. Standing topics:
  4. Anyone have experience with pure column configurations?

Notes:

Participants: Adrian, Matt, Mark

Apologies to anyone in the CET timezone who missed the call because Michael is not good with time zones and daylight savings time! Next week's telecon will start at 1800 CET

  • Logging module: next step would be to convert all cores in 'develop'

    • start with framework (any complications should turn up there)
    • maybe at some point we need to revisit developer guidelines (e.g., should every commit compile)
    • for the conversion, don't leave in old write statements -- we can always compare before and after branches
    • changes to sea-ice column package that would be needed for logging module have just been merged
    • Mark: what about timeline for individual cores? may take until after spring break
    • if timeline is too long, other merges into develop may cause conflicts
    • maybe it would be good if we all worked on this the same week: next week 20-24 March - next telecon we'll get instructions
    • maybe come up with a script or vim macro

  • Framework PRs:

    • 1255 Adrian will update the initial comment, Michael will merge
    • 1237 a test with ACME isn't coming out right - more investigation needed
    • 1227 not immediately needed, see whether there is a more general solution?

  • Subset I/O capability:

    • Matt: probably better to do this on a per-stream basis, need self-contained output without a level of indirection to get, e.g., x/y/zCell for elements as this would be problematic for visualization software
    • Mark: scope would be simplified if all mesh fields could be pre-processed and associated with subsets of fields during postprocessing; all that MPAS would need to do is output the correct subset of field elements
    • An initial document will be put together for the next telecon

  • Standing topics:

    • Merge of CIME 5.2 is still in progress

  • Single-column configurations: does anyone have experience with this?

    • MPAS-Ocean has been run with 4x4, doubly periodic configuration
    • MPAS-Atmosphere has been run with 2xn doubly periodic configuration
    • Current mesh-generation software probably won't work with a single column
    • Adrian may experiment with code to create a single-column mesh
      • Can we have cellsOnCell field for a cell refer to the cell itself?
Clone this wiki locally