Nexus doesn't detect run failure #5833
Description
Describe the bug
Running lab2_qmc_basics/oxygen_dimer without QE binaries on the PATH.
Modify machine to ws8.
$ python O_dimer.py
...
starting runs:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
elapsed time 0.0 s memory 105.95 MB
Entering ./scale_1.0 0
writing input files 0 dft
Entering ./scale_1.0 0
sending required files 0 dft
submitting job 0 dft
Entering ./scale_1.0 0
Executing:
export OMP_NUM_THREADS=1
mpirun -np 4 pw.x -input dft.in
elapsed time 3.0 s memory 105.98 MB
elapsed time 6.1 s memory 105.99 MB
elapsed time 9.1 s memory 105.99 MB
No running pw.x and no error caught. nexus just sits idle.
Then export PATH=my_qe_installation/bin:$PATH
$ python O_dimer.py
...
starting runs:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
elapsed time 0.0 s memory 105.98 MB
elapsed time 3.0 s memory 105.98 MB
elapsed time 6.1 s memory 105.99 MB
elapsed time 9.1 s memory 105.99 MB
elapsed time 12.1 s memory 105.99 MB
pw.x still doesn't run and no error caught. nexus just sits idle.
$ rm scale_1.0
$ python O_dimer.py
...
starting runs:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
elapsed time 0.0 s memory 106.28 MB
Entering ./scale_1.0 0
writing input files 0 dft
Entering ./scale_1.0 0
sending required files 0 dft
submitting job 0 dft
Entering ./scale_1.0 0
Executing:
export OMP_NUM_THREADS=1
mpirun -np 4 pw.x -input dft.in
elapsed time 3.0 s memory 577.72 MB
elapsed time 6.1 s memory 2247.41 MB
elapsed time 9.1 s memory 2919.17 MB
elapsed time 12.2 s memory 3127.30 MB
finally it runs. then next is qmcpack run
Entering ./scale_1.0 2
writing input files 2 opt
Entering ./scale_1.0 2
sending required files 2 opt
submitting job 2 opt
Entering ./scale_1.0 2
Executing:
export OMP_NUM_THREADS=1
mpirun -np 4 qmcpack opt.in.xml
elapsed time 134.7 s memory 106.50 MB
elapsed time 137.7 s memory 106.51 MB
elapsed time 140.8 s memory 106.51 MB
elapsed time 143.8 s memory 106.51 MB
again it is idle again dule to missing qmcpack executables on the PATH.
Expected behavior
Properly error out when mpirun fails. Probably also worth checking the existence of executables.
System:
- laptop ws8