No solution.pyfrs

57 views
Skip to first unread message

Liam Doult

unread,
Dec 7, 2016, 7:03:05 AM12/7/16
to PyFR Mailing List
Hi, 

my current error: 

/scratch/benchmarks/pyfr/bin/pyfr partition 8 cube_hex24.pyfrm /scratch/benchmarks/pyfr/cube_hex/solution.pyfrs .
Traceback (most recent call last):
  File "/scratch/benchmarks/pyfr/bin/pyfr", line 11, in <module>
    load_entry_point('pyfr==1.5.0', 'console_scripts', 'pyfr')()
  File "/scratch/benchmarks/pyfr/lib/python3.5/site-packages/pyfr-1.5.0-py3.5.egg/pyfr/__main__.py", line 110, in main
  File "/scratch/benchmarks/pyfr/lib/python3.5/site-packages/pyfr-1.5.0-py3.5.egg/pyfr/__main__.py", line 171, in process_partition
  File "/scratch/benchmarks/pyfr/lib/python3.5/site-packages/pyfr-1.5.0-py3.5.egg/pyfr/__main__.py", line 164, in <genexpr>
  File "/scratch/benchmarks/pyfr/lib/python3.5/site-packages/pyfr-1.5.0-py3.5.egg/pyfr/readers/native.py", line 16, in __init__
  File "/usr/lib64/python3.5/site-packages/h5py/_hl/files.py", line 272, in __init__
    fid = make_fid(name, mode, userblock_size, fapl, swmr=swmr)
  File "/usr/lib64/python3.5/site-packages/h5py/_hl/files.py", line 92, in make_fid
    fid = h5f.open(name, flags, fapl=fapl)
  File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper (/tmp/pip-at6d2npe-build/h5py/_objects.c:2684)
  File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper (/tmp/pip-at6d2npe-build/h5py/_objects.c:2642)
  File "h5py/h5f.pyx", line 76, in h5py.h5f.open (/tmp/pip-at6d2npe-build/h5py/h5f.c:1930)
OSError: Unable to open file (Unable to open file: name = '/scratch/benchmarks/pyfr/cube_hex/solution.pyfrs', errno = 2, error message = 'no such file or directory', flags = 0, o_flags = 0)

I managed to get the application to run with 
/scratch/benchmarks/pyfr/bin/pyfr partition 8 cube_hex24.pyfrm .
But this is not producing any output file.

my mpirun: 
mpirun -np 10 --map-by node -machinefile machinefile -x PATH -x LD_LIBRARY_PATH -x OMP_NUM_THREADS=1 -x PYTHONPATH /scratch/benchmarks/pyfr/bin/pyfr run -b openmp -p cube_hex24.pyfrm  config.ini

any solutions?
Please?

Regards
Liam

Ray

unread,
Dec 7, 2016, 7:11:50 AM12/7/16
to PyFR Mailing List
Hi, Liam, I think you should use this command:
pyfr partition 8 cube_hex24.pyfrm "path"

"path" is a  folder at your current folder. Then you'll see cube_hex24.pyfrm in folder "path".


在 2016年12月7日星期三 UTC+8下午8:03:05,Liam Doult写道:

Vincent, Peter E

unread,
Dec 7, 2016, 7:41:11 AM12/7/16
to PyFR Mailing List
Hi Liam,

/scratch/benchmarks/pyfr/bin/pyfr partition 8 cube_hex24.pyfrm .
But this is not producing any output file.

As per Ray’s message, the last entry is the path to where you want the partitioned mesh to be put. Since you have put ‘.' the partitioned mesh will be written to the same directory as the original mesh, and as such will overwrite it (hence you do not see a new file appear, but cube_hex24.pyfrm should now have 8 partitions).

mpirun -np 10 --map-by node -machinefile machinefile -x PATH -x LD_LIBRARY_PATH -x OMP_NUM_THREADS=1 -x PYTHONPATH /scratch/benchmarks/pyfr/bin/pyfr run -b openmp -p cube_hex24.pyfrm  config.ini

It appears you are partitioning the mesh into 8 pieces, but launching 10 MPI ranks?

Cheers

Peter

Dr Peter Vincent MSci ARCS DIC PhD
Reader in Aeronautics and EPSRC Fellow
Department of Aeronautics
Imperial College London
South Kensington
London
SW7 2AZ
UK

web: www.imperial.ac.uk/aeronautics/research/vincentlab
twitter: @Vincent_Lab





--
You received this message because you are subscribed to the Google Groups "PyFR Mailing List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pyfrmailingli...@googlegroups.com.
To post to this group, send email to pyfrmai...@googlegroups.com.
Visit this group at https://groups.google.com/group/pyfrmailinglist.
For more options, visit https://groups.google.com/d/optout.

Liam Doult

unread,
Dec 7, 2016, 7:56:34 AM12/7/16
to PyFR Mailing List
Hi, thanks for the quick response

-np 10 is a typo but not the problem. The problem is generating the .pyfrs file with data.

regards 
Liam

Niki Loppi

unread,
Dec 7, 2016, 8:46:14 AM12/7/16
to pyfrmai...@googlegroups.com

Hi Liam,

Have you created the /scratch/benchmarks/pyfr/cube_hex/  folder for output?

Just specifying the 'basedir=cube_hex' in the .ini file does not create it, so you have to do it manually via mkdir.

Thanks,

Niki

--
You received this message because you are subscribed to the Google Groups "PyFR Mailing List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pyfrmailingli...@googlegroups.com.
To post to this group, send email to pyfrmai...@googlegroups.com.
Visit this group at https://groups.google.com/group/pyfrmailinglist.
For more options, visit https://groups.google.com/d/optout.

-- 
Niki Andreas Loppi MSc
Postgraduate Researcher
Department of Aeronautics
Imperial College London
South Kensington
London
SW7 2AZ
UK

Freddie Witherden

unread,
Dec 7, 2016, 9:20:13 AM12/7/16
to pyfrmai...@googlegroups.com
Hi Liam,

On 07/12/2016 04:03, Liam Doult wrote:
> I managed to get the application to run with
> /scratch/benchmarks/pyfr/bin/pyfr partition 8 cube_hex24.pyfrm .
> But this is not producing any output file.
>
> my mpirun:
> mpirun -np 10 --map-by node -machinefile machinefile -x PATH -x
> LD_LIBRARY_PATH -x OMP_NUM_THREADS=1 -x PYTHONPATH
> /scratch/benchmarks/pyfr/bin/pyfr run -b openmp -p cube_hex24.pyfrm
> config.ini

It is difficult to know what is going on without the configuration file;
after all that is where one tells PyFR how often files should be written
out to disk.

However, my psychic debugging skills indicate that config.ini is from a
previous version of PyFR that used a different syntax for specifying
output files.

For PyFR v.1.50 you need to ensure that there is a section entitled
[soln-plugin-writer] as per the user guide.

Regards, Freddie.


signature.asc

Liam Doult

unread,
Dec 7, 2016, 10:27:49 AM12/7/16
to PyFR Mailing List
Thank you for you input Niki. That was the problem.

Regards
Liam

On Wednesday, December 7, 2016 at 2:03:05 PM UTC+2, Liam Doult wrote:
Reply all
Reply to author
Forward
0 new messages