Is your feature request related to a problem?
spack install can take a very long time if run serially with only very few build jobs. It does provide good parallelism, especially since spack v1, but this can't be used on login nodes for HPCs.
For NRL, we need to be able to offload the spack install step to the compute nodes on all the machines.
Describe the solution you'd like
Implement a feature (switch, command line argument) to offload the spack install step to the compute nodes using the batch scheduler, and make use of the parallelism that spack install provides.
Additional context
n/a