Hi everyone,
I am currently trying to run a two-domains one-way nested simulation run with WRF v4.3 and using ERA5 as forcing. My simulation runs just fine when only using one node with 32 procs (max nprocs per node). However, when I increase the number of nodes (e.g. 2 nodes), model crashes at the first timestep. I have tried using 2 nodes with varying numbers of procs each (e.g., 32, 24, 16, 2) without success. However, when using one single node with as little as 2 procs (I know this should not be done because of my domains size), the model starts solving the timesteps just fine. I must also say that these errors do not happen when running a single domain simulation, in this case I can use more than one node without any problem.
Domain 1 is 301 x 301 and Domain 2 is 301 x 301 as well.
What can be done to fix this?
Attached files are a figure of the domains, my namelist.input and the 64 rsl.errors and 64 rsl.outs zipped in a tar file
Cheers,
Ricardo
I am currently trying to run a two-domains one-way nested simulation run with WRF v4.3 and using ERA5 as forcing. My simulation runs just fine when only using one node with 32 procs (max nprocs per node). However, when I increase the number of nodes (e.g. 2 nodes), model crashes at the first timestep. I have tried using 2 nodes with varying numbers of procs each (e.g., 32, 24, 16, 2) without success. However, when using one single node with as little as 2 procs (I know this should not be done because of my domains size), the model starts solving the timesteps just fine. I must also say that these errors do not happen when running a single domain simulation, in this case I can use more than one node without any problem.
Domain 1 is 301 x 301 and Domain 2 is 301 x 301 as well.
What can be done to fix this?
Attached files are a figure of the domains, my namelist.input and the 64 rsl.errors and 64 rsl.outs zipped in a tar file
Cheers,
Ricardo
