Upstream old fixes / features

[henryleberre]

I realized that I forgot to upstream these fixes / features:

• Enable testing on multiple GPUs (in parallel).
• The above point and other changes make it easier to implement benchmarking in the future.
• Small build.py refactor. This also fixes a bug relating to targets not being built in a deterministic order. I understood this would happen when I pushed those changes but underestimated how confusing that might be.

[sbryngelson]

Will merge this once tests pass -- was failing before for some spurious thing that is now fixed.

[sbryngelson]

[SUCCESS]
++ ok 'All modules have been loaded for GT Phoenix on GPUs.'
++ log 'OK > All modules have been loaded for GT Phoenix on GPUs.'
++ echo -e 'mfc: OK > All modules have been loaded for GT Phoenix on GPUs.'
mfc: OK > All modules have been loaded for GT Phoenix on GPUs.
++ return
++ nvidia-smi -L
++ wc -l
+ gpu_count=2
++ seq -s , 0 1
+ gpu_ids=0,1
++ nproc
+ ./mfc.sh test -a -b mpirun -j 24 --gpu -g 0,1
mfc: OK > (venv) Entered the Python virtual environment.
usage: ./mfc.sh test [-h] [--mpi] [--no-mpi] [--gpu] [--no-gpu] [--debug]
                     [--no-debug] [-j JOBS] [-v] [--no-fftw] [--no-hdf5]
                     [--no-silo] [-g GPUS [GPUS ...]] [-l] [-f FROM] [-t TO]
                     [-o L [L ...]] [-b {jsrun,srun,mpirun,mpiexec,N/A}] [-r]
                     [-a] [-% PERCENT] [-m MAX_ATTEMPTS] [--case-optimization]
                     [--generate | --add-new-variables]
./mfc.sh test: error: argument -g/--gpus: invalid int value: '0,1'
mfc: (venv) Exiting the Python virtual environment.
---------------------------------------
Begin Slurm Epilog: Oct-08-2023 12:54:15
Job ID:        3730688
Array Job ID:  _4294967294
User ID:       sbryngelson3
Account:       gts-sbryngelson3
Job name:      shb-test-jobs
Resources:     cpu=24,gres/gpu:v100=2,mem=96G,node=1
Rsrc Used:     cput=00:02:00,vmem=[165](https://github.com/MFlowCode/MFC/actions/runs/6443848089/job/17506330869?pr=218#step:5:166)6K,walltime=00:00:05,mem=0,energy_used=0
Partition:     gpu-v100
QOS:           embers
Nodes:         atl1-1-01-003-36-0
---------------------------------------