handle CPU or GPU via flag? right now using Delta on CPU breaks if CuPy is installed

[lastephey]

Delta right now imports CuPy whether it's running on the CPU or GPU. This is a problem when running on the CPU but in the Delta container where CuPy is installed:

stephey@nid00218:/global/cscratch1/sd/stephey/delta/delta> OMP_NUM_THREADS=16 srun -n 16 -c 2 --cpu-bind=cores shifter python3 -m mpi4py.futures processor.py --config configs/hackathon_test.json --transport transport_tx  --num_ranks_preprocess=1 --num_ranks_analysis=16 --num_queue_threads=1 --run_id=test_25259_GT
Traceback (most recent call last):
  File "/usr/local/lib/python3.8/dist-packages/cupy/__init__.py", line 20, in <module>
    from cupy import core  # NOQA
  File "/usr/local/lib/python3.8/dist-packages/cupy/core/__init__.py", line 1, in <module>
    from cupy.core import core  # NOQA
ImportError: libcuda.so.1: cannot open shared object file: No such file or directory

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/usr/lib/python3.8/runpy.py", line 194, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/usr/lib/python3.8/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/usr/local/lib/python3.8/dist-packages/mpi4py/futures/__main__.py", line 72, in <module>
    main()
  File "/usr/local/lib/python3.8/dist-packages/mpi4py/futures/__main__.py", line 60, in main
    run_command_line()
  File "/usr/local/lib/python3.8/dist-packages/mpi4py/run.py", line 47, in run_command_line
    run_path(sys.argv[0], run_name='__main__')
  File "/usr/lib/python3.8/runpy.py", line 265, in run_path
    return _run_module_code(code, init_globals, run_name,
  File "/usr/lib/python3.8/runpy.py", line 97, in _run_module_code
    _run_code(code, mod_globals, init_globals,
  File "/usr/lib/python3.8/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "processor.py", line 28, in <module>
    from analysis.task_list import tasklist
  File "/global/cscratch1/sd/stephey/delta/delta/analysis/task_list.py", line 6, in <module>
    from analysis.helpers import get_analysis_task
  File "/global/cscratch1/sd/stephey/delta/delta/analysis/helpers.py", line 10, in <module>
    from analysis.task_spectral_cu import task_coherence_cu, task_crosscorr_cu, task_crossphase_cu, task_crosspower_cu
  File "/global/cscratch1/sd/stephey/delta/delta/analysis/task_spectral_cu.py", line 4, in <module>
    from analysis.kernels_spectral_cu import kernel_crosscorr_cu, kernel_coherence_cu, kernel_crosspower_cu, kernel_crossphase_cu
  File "/global/cscratch1/sd/stephey/delta/delta/analysis/kernels_spectral_cu.py", line 7, in <module>
    import cupy as cp
  File "/usr/local/lib/python3.8/dist-packages/cupy/__init__.py", line 41, in <module>
    raise ImportError(_msg) from e
ImportError: CuPy is not correctly installed.

I suggest that Delta have an --arch or --mode flag that a user can set to CPU or GPU depending on what they need. Then Delta could handle the imports accordingly.