[Ncep.list.wwatch3.discussion-group] WWIII Notes - Weds, Sept 26, 2018

[Jessica Meixner - NOAA Federal]

WWIII Notes - Weds, Sept 26, 2018

Hi Everyone -

Thank you all for those who joined us today.  The notes from the meeting
are below, but first the highlights:


   -

   The deadline for new features will be Wednesday, October 17th for the
   next public release.
   -

   On October 17th, we will be making a public release branch which we will
   begin testing.  At that point only bug fixes and documentation will be
   accepted on this branch.
   -

   Please plan to have all documentation updates by November 1st.
   -

   The next public release will be via github (no more
   registration/password firewalls).
   -

   We will be looking into moving to open development via github in the
   future using gitflow:
   https://nvie.com/posts/a-successful-git-branching-model/
   -

   Next meeting is Wednesday, October 17th at 10:30am.


More details, that might require a response:

   -

   We need to rename the aux directory. (Sorry I skipped this in the call).
   I propose to rename this to auxiliary. You can reply here or directly on
   this issues https://vlab.ncep.noaa.gov/redmine/issues/46439 by selecting
   “edit” and then adding your comments to the notes.
   -

   See Mickael’s email about using comp/link.tmpl and test this out.
   Hopefully this will help us clean up and remove extra comp/link things.
   -

   Please provide input to Andy’s proposal to update partition and TM10
   variable names.  You can reply here or directly on the issue:
   https://vlab.ncep.noaa.gov/redmine/issues/55541 by selecting “edit” and
   then adding your comment in the notes.


Status for public release:

   -

   Update on codes ready to be merged back in:   ESMF unstructured grids is
   working and just needs to be pushed back to the master, BToutput has a
   small issue we are working through but is otherwise ready.
   -

   For namelists, ww3_grid.nml should be ready for the next public
   release.  There are many programs that still do not have namelist options
   that will not be ready for the public release.  If you have a high priority
   one, please let Mickael know and he will try to have that ready as well.
   -

   Bit reproducibility update:
   -

      For the goal of having the NCEP operational set-up be bit
      reproducible (See ticket:
      https://vlab.ncep.noaa.gov/redmine/issues/45652) the issue was solved
      by redesigning the grid set-up. This bypassed an issue of bit
      reproducibility when having two grids of the same rank.  By adding MPI
      barriers, this issue would go away in the beginning of calculations but
      there was a performance hit. Henrique will add more details to
the issue to
      summarize his findings.
      -

      We still have bit reproducibility issues for the regtests in
      general.  To see a summary see:
      https://vlab.ncep.noaa.gov/redmine/projects/emc-ww3/wiki/How_to_use_matrixcomp_to_compare_regtests_with_master
       It would be great to reduce the number of regtests that are not
      bit-for-bit.  Any help is greatly appreciated.
      -

      There was a follow on to this topic about the need for same-rank
      grids and other advantages to particular strategies of grid
design. Erick &
      Henrique are going to follow up offline and will report back findings to
      the whole group.
      -

   Other updates will be pushed to the next public release.


General code management updates:

   -

   If you have branches that can be deleted please let Ali or I know.
   -

   On the wiki page (
   https://vlab.ncep.noaa.gov/redmine/projects/emc-ww3/wiki#Upcoming-commits)
   there is now a section called “upcoming commits” which has a list of all
   the branches in line/ready to be tested to go back to the master.
   -

   To submit to the master, email code-managers with name of your branch &
   answers from here
   <https://vlab.ncep.noaa.gov/redmine/projects/emc-ww3/wiki/Checklist_for_submitting_to_master>
    instead of pushing to code-review directly.
   -

   Due to operational requirements, we must remove all computed GOTOs from
   WW3 and reduce the overall number.  So please do not add GOTOs in
   additional code.


   -

   The bin/ directory is in need of cleaning before the public release.
   Please look at your switch files, the comp/link (see Mickael’s email
   first), and the READMEs for updates.
   -

   There are lots of outstanding issues, to see those assigned to you:
   https://vlab.ncep.noaa.gov/redmine/projects/emc-ww3/issues?query_id=259
    and if you want to help solve another issue - thank you! - just let the
   person it’s “assigned” to know.


If I missed something, please add it in and as always if you have questions
let Ali or I know.

Best,

Jessica


-- 
Jessica Meixner
NCEP/EMC
NOAA Center for Weather and Climate Prediction
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://www.lstsrv.ncep.noaa.gov/pipermail/ncep.list.wwatch3.discussion-group/attachments/20180926/398f1cb1/attachment-0001.html