Hi all,
I ran my CTSM/CLM case for one simulation year, and the runtime with 120 MPI tasks was about 3 hours and 30 minutes. The case covers the northern domain from 25°N to 90°N.
I was wondering whether it would be reasonable to reduce the runtime by splitting the full 25°N–90°N domain into smaller latitude bands:
and submitting the separate regional cases at the same time.
However, I am not fully sure this is a good solution. As I understand it, I would need to create separate regional cases and prepare separate subsetted input/domain files for each latitude band. This may take some time, and the outputs will be split into separate files rather than a single complete file for the full domain.
Could you please advise whether this latitude-band approach is actually recommended ?
thyank you in advance for your guidance.
regards,
shahla
I ran my CTSM/CLM case for one simulation year, and the runtime with 120 MPI tasks was about 3 hours and 30 minutes. The case covers the northern domain from 25°N to 90°N.
I was wondering whether it would be reasonable to reduce the runtime by splitting the full 25°N–90°N domain into smaller latitude bands:
and submitting the separate regional cases at the same time.However, I am not fully sure this is a good solution. As I understand it, I would need to create separate regional cases and prepare separate subsetted input/domain files for each latitude band. This may take some time, and the outputs will be split into separate files rather than a single complete file for the full domain.
Could you please advise whether this latitude-band approach is actually recommended ?
thyank you in advance for your guidance.
regards,
shahla