Making our mkl compile config work directly in the compile script rather than relying on a correspondence with the config files. Also settig main config scripts to use new ifx compiler. (details)
Removing allocatable status of some variables. These need not be declared allocatable and ifx seems to have a problem with it. The allocatable status is an artifact from when we had these as arrays of initially unknown size. (details)
Fixing the run agent to be carson, since it has the only GPU capable of running with openmp gpu offloading. (details)
* Initial changes to get code working with the new spack+module system. See https://github.com/larson-group/sys_admin/issues/829
* Changes for restart_gfortran_test
* Possible fix for tuning test
* Part 2 to tuning test fix. It's unclear why the recent change caused top level directory for these stages to be in clubb, rather than the parent directory. Perhaps it has to do with the triple quote
* Adding the mkl flag back to intel config scripts. It's still usable when adding -m or --mkl_lapack to our compile script call, and now we have intel-oneapi-mkl installed because wrf needs it.
Making our mkl compile config work directly in the compile script rather than relying on a correspondence with the config files. Also settig main config scripts to use new ifx compiler.
Removing allocatable status of some variables. These need not be declared allocatable and ifx seems to have a problem with it. The allocatable status is an artifact from when we had these as arrays of initially unknown size.