Runtime error about 2d euler vortex example

General
1

A few days ago, i update pyfr version (1.13 → 1.14.0)
Thereafter the 2D euler vortex example is not executed.
----------------------------error message--------------------

(pyfr_14) sk@sk:~/pyfr/pyfr_14/PyFR-1.14.0/examples/euler_vortex_2d$ mpiexec -n 2 pyfr run -b cuda -p euler_vortex_2d.pyfrm euler_vortex_2d.ini
Traceback (most recent call last):
File “/home/sk/pyfr/pyfr_14/bin/pyfr”, line 33, in
sys.exit(load_entry_point(‘pyfr==1.14.0’, ‘console_scripts’, ‘pyfr’)())
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/main.py”, line 115, in main
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/main.py”, line 248, in process_run
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/main.py”, line 230, in _process_common
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/rank_allocator.py”, line 14, in get_rank_allocation
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/rank_allocator.py”, line 31, in init
RuntimeError: Mesh has 2 partitions but running with 1 MPI ranks
Traceback (most recent call last):
File “/home/sk/pyfr/pyfr_14/bin/pyfr”, line 33, in
sys.exit(load_entry_point(‘pyfr==1.14.0’, ‘console_scripts’, ‘pyfr’)())
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/main.py”, line 115, in main
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/main.py”, line 248, in process_run
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/main.py”, line 230, in _process_common
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/rank_allocator.py”, line 14, in get_rank_allocation
File “/home/sk/pyfr/pyfr_14/lib/python3.9/site-packages/pyfr-1.14.0-py3.9.egg/pyfr/rank_allocator.py”, line 31, in init
RuntimeError: Mesh has 2 partitions but running with 1 MPI ranks

I think it is an error related to MPI parallelization
What is the cause of the error and how can it be resolved?

2

If seems as though mpi is running the 2 instances of the pyfr command rather than pyfr using two mpi ranks.

Did you also update mpi recently? If so you will need to rebuild mpi4py.

3

That problem was fixed. Thanks Will !

Additionally, I will build new computer to proceed GPU computing using 4 GPUs.
Can i run PyFR in 1 CPU (8 cores) + 4 GPUs (Nvidia) ?

4

Yes, if you run PyFR with a command something like mpiexec -n 4 pyfr run -b cuda ... that should do the trick. If you haven’t already it’s probably worth compiling MPI with the CUDA aware flag, see here for more details: Performance Tuning — Documentation