Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

Clarification on run type for restarting a run

rambhari01

Ram
New Member
Dear CESM users,

I need some clarification on deciding the run type for 2 set of experiments. For both of these experiment I want to use the restart file from a run say a control run X.

1.) I want to continue the run X with exactly same specifications but changed some of resource/machine related things such as (number of processor/nodes etc.). Since, everything is going to be same with run X, Can i select the run_type='branch' for this and follow the guidelines for starting a branch run? as per my understanding this will continues the run X from the restart check points?

2.) for the second type of run, I want to use the restart files from the check point from the run X stopped. But for this new run, I want to make several changes such as
a.) a compset type but for similar type of experiment (such as F-compset for 1850 and hist, so that everything related with the model components is consistent between X and this run).
b.) I also want to change some input files, such as Findata, surfdata etc and namelist variable controlling the CO2 to diagnostic and fixed etc.
c.) also, I want to change the start dates for this new run. but want the model initialization from the X (restart files).

For this category of runs, I assume that I should follow guidelines for a 'hybrid' run by setting the run_type='hybrid'.

Please clarify if my understanding the run_types (branch and hybrid) is okay and I will be able to start those 2 desired set of run using the restart files from control run X.

Also, please correct or mention if I missed something or stated above incorrectly.

Thank you very much.

-Ram
 

sacks

Bill Sacks
CSEG and Liaisons
Staff member
For (1) yes, a branch run should work. If you're just changing the processor count, you should also be able to do this by setting CONTINUE_RUN=TRUE (then run case.setup --reset followed by case.build), which could be simpler than creating a new branch case.

For (2) yes, a hybrid case is appropriate. Depending on how significant the changes are from the initial run, even a hybrid case may fail to start up, but as long as your new run is pretty consistent with your original, then a hybrid will work.
 
Vote Upvote 0 Downvote
Top