Sum of the atmosphere sub-timers is greater than ATM_RUN, so we used the sum of timers for the stacked plot.
This happens in certain cases where MPI wallmax timers are different across processors. Typically it's harmless if it is a small percentage.
This issue cause the total percent to be over 100%.
Sum of the atmosphere sub-timers is greater than ATM_RUN, so we used the sum of timers for the stacked plot. This happens in certain cases where MPI wallmax timers are different across processors. Typically it's harmless if it is a small percentage.
This issue cause the total percent to be over 100%.
add this to the unified plot/plain atm as well