You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
During testing of #1005, I discovered that the seed dispersal mechanism is not b4b between different layout changes (i.e the test runs fail ERP or PEM tests). This is likely due to order of operation issues in collecting floating point values with the MPI collective routines. Note that ERS tests pass COMPARE_BASE_REST checks.
The text was updated successfully, but these errors were encountered:
Note that ESCOMP/CTSM#2077 adds a cross-grid seed dispersal testmod that is called with a layout changing PEM testtype option. This will further help us track and address the issue in the future.
During testing of #1005, I discovered that the seed dispersal mechanism is not b4b between different layout changes (i.e the test runs fail ERP or PEM tests). This is likely due to order of operation issues in collecting floating point values with the MPI collective routines. Note that ERS tests pass COMPARE_BASE_REST checks.
The text was updated successfully, but these errors were encountered: