Agenda Notes Monday February 13th 2017 11:00 AM

Date: 13 February 2017
Time: 11:00 MST / 19:00 CET

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

Participants: Dom, Mark, Matt, Jon, Adrian

1237
All cores should first check that there is no harm done
When testing with MPI only, expect no change in performance
Expect there may be performance improvements on Phi/KNL with OpenMP
Please post any feedback from testing, especially timing results
Dom will test with alternative OpenMP implementation in MPAS-Atmosphere
1232
Basic compilation test should be all that’s needed from each core
Will need to feed these changes into the CIME repository
1227
Problem in framework forcing module
cycling the forcing clock causes times to get out of sync
calculation of interpolation weights -> sum to 1, but may be negative -> supersonic winds
would it work to fix mpas_calibrate_alarms so that it works when time is set to something in the past?
Michael: interested to rework mpas_calibrate_alarms anyway
Adrian: can make a test case to demonstrate the problem
1211
See comments
When PIO2 eventually supports fill value, and ACME adopts that version, we’ll update MPAS
Matt will update the branch later today, then should be ready to merge
ACME team talking about adding fill values to PIO 1.x

Provide feedback