Error when partitioning test example 2d Euler

Errors
1

I received the no partitioners error as I’ve seen others have in the forums. I had metis installed but it is version 5.1 - I see comments that we should be using 5.2.

Unfortunately I cannot seem to find this version anywhere? Any guidance on getting this installed (or the alternative) would be appreciated.

I am using WSL ubuntu.

2

Regards, Freddie.

3

I tried that release but still had trouble.

I attempted a fresh install following the ubuntu installation guide. I am using WSL.

All installed fine. I then got to the 2d euler vortex example where it failed at step 4, giving me this error:

> mpiexec -n 2 pyfr -p run -b cuda euler-vortex.pyfrm euler-vortex.ini
Traceback (most recent call last):
  File "/path/to/pyfr-venv/bin/pyfr", line 8, in <module>
    sys.exit(main())
             ^^^^^^
  File "/path/to/pyfr-venv/lib/python3.11/site-packages/pyfr/__main__.py", line 124, in main
    args.process(args)
  File "/path/to/pyfr-venv/lib/python3.11/site-packages/pyfr/__main__.py", line 258, in process_run
    _process_common(
  File "/path/to/pyfr-venv/lib/python3.11/site-packages/pyfr/__main__.py", line 237, in _process_common
    backend = get_backend(args.backend, cfg)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/pyfr-venv/lib/python3.11/site-packages/pyfr/backends/__init__.py", line 11, in get_backend
    return subclass_where(BaseBackend, name=name.lower())(cfg)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/path/to/pyfr-venv/lib/python3.11/site-packages/pyfr/backends/cuda/base.py", line 18, in __init__
    self.cuda = CUDA()
                ^^^^^^
  File "/path/to/pyfr-venv/lib/python3.11/site-packages/pyfr/backends/cuda/driver.py", line 444, in __init__
    self.lib.cuInit(0)
  File "/path/to/pyfr-venv/lib/python3.11/site-packages/pyfr/ctypesutil.py", line 37, in _errcheck
    raise self._statuses[status]
pyfr.backends.cuda.driver.CUDANoDevice
--------------------------------------------------------------------------
prterun detected that one or more processes exited with non-zero status,
thus causing the job to be terminated. The first process to do so was:

   Process name: [prterun-hostname-138003@1,1]
   Exit code:    1

I ran nvidia-smi and nvcc --version which showed my gpu and driver details.

Any further help appreciated.

4

Does it work if you try and run with only a single partition?

Regards, Freddie.

5

Thanks Freddie.

I had to run

pyfr partition 1 euler-vortex.pyfrm .

and then

mpiexec -n 1 pyfr -p run -b cuda ./euler-vortex.pyfrm euler-vortex.ini

and then the pyfrs files generated.