scream
scream copied to clipboard
E3SM-Project
Exascale global atmosphere model written in C++ as part of the E3SM project
There are a number of issues that come up when running a CIME based case w/ nudging turned on. This issue lists them so we can track progress on addressing...
There are some fields computed by rad that are only relevant in certain areas (e.g., on cols where it's day), and that are set to something like "fillValue" elsewhere. When...
In https://github.com/E3SM-Project/scream/issues/2391, we hit a NaN after 35 days. I wanted to see if using DEBUG build and continuing would give us any more info about source of error, however,...
I'm currently hitting what looks like a hang on tioga with ne4 and 1 MPI/GPU. `SMS_D_P1x1_Ld1.ne4pg2_ne4pg2.F2010-SCREAMv1.tioga_craygpu` I can see stack with gdb ``` #0 0x000015553f066868 in rocr::core::InterruptSignal::WaitRelaxed(hsa_signal_condition_t, long, unsigned long,...
For a "clean" build, ie using `create_test`, I see this link line: ``` python3 /p/lustre1/keen3/e3sm_scratch/tioga/s05-may3/SMS_D_P1x1_Ld1.ne4pg2_ne4pg2.F2010-SCREAMv1.tioga_craygpu.may25b/Tools/e3sm_compile_wrap.py /opt/cray/pe/craype/2.7.20/bin/ftn -Wl,--allow-multiple-definition -L/opt\ /rocm-5.4.3/lib -hsystem_alloc -I/opt/cray/pe/mpich/8.1.25/ofi/crayclang/10.0/include "CMakeFiles/e3sm.exe.dir/p/lustre1/keen3/wacmy/s05-may3/driver-mct/main/cime_comp_mod.F90.o" "CMakeFiles/e3sm.exe.dir/p/lustre1/keen3/\ wacmy/s05-may3/driver-mct/main/cime_driver.F90.o" "CMakeFiles/e3sm.exe.dir/p/lustre1/keen3/wacmy/s05-may3/driver-mct/main/component_mod.F90.o" "CMakeFiles/e3sm.exe.dir/p/lustre1/keen3/wacmy/s05-may3/driver-mct/main/compo\ nent_type_mod.F90.o" "CMakeFiles/e3sm.exe.dir/p/lustre1/keen3/wacmy/s05-may3/driver-mct/main/cplcomp_exchange_mod.F90.o" "CMakeFiles/e3sm.exe.dir/p/lustre1/keen3/wacmy/s05-may3/driver-mct/main/map_glc2lnd_mod.F90.o" "CMa\...
`p_mid` is removed from nudging. However, in atmosphere nudging, we still need to use the internal `p_mid` to `perform_vertical_interpolation`.
It was suggested to document Frontier instabilities here. In preparation for production runs, we submitted 10 512-node runs for ne1024pg2. 5 runs succeeded, 5 runs fails due to different reasons....
I occasionally see (various) jobs hang on pm (pm-cpu or pm-gpu) and I've been trying to debug in general, but this one seems different. If I run a basic ne30...
For CPU builds, YAKL/radiation is not using threading. We need to at least build with a variable to enable this. I will try manually testing. It was suggested that we...
Some processes make use of internal `helper` fields. It would be nice to have an option to flush these to an output. It would make it easier to track how...
Metadata
Owner
Metadata
Exascale global atmosphere model written in C++ as part of the E3SM project