NUOPC's ESMF-aware threading gives incorrect resource line to job system (when using cores as threads)

When running CESM with ESMF_AWARE_THREADING=true, the 'ompthreads' line would specify the number of threads desired, vs letting ESMF's managed threading handle them. This results in ESMF PETs being assigned strided CPU cores (by the number of threads specified), and no ability for it to 'map' those skipped CPU cores into its threading logic.

It might be this was desired behavior with 'hyperthreads' (eg, map to a single core, but two hyperthreads on it), but is not desired behavior for full core-based threading.