Post-Processing using LPP

Theo_Score's picture
Submitted by Theo_Score on Mon, 02/20/2017 - 04:45

Hi All,

I meant to ask if it is possible to include a command in the job input script to perform post-processing.

For instance, I have included in my script;

variable a universe 1 2 3 4 5 6 7 8 9

and defined results post directories as;

post_0.$a

My dump files have the format dump*.data so;

shell cd post_0.$a (I am sure gets me into the directories) and now I want to execute;

lpp dump*.data (Which I normally execute from the terminal - How do I include this in my job input script).

Thank you.

Forums:

Daniel Queteschiner | Mon, 02/20/2017 - 09:20

LIGGGHTS might be unable to resolve the lpp alias, so just type the full command, i.e.
shell cd post
shell python /path/to/lpp/src/lpp.py dump*.data
However I recommend using dump custom/vtk instead to save the trouble of post-processing.

Cheers
Daniel

Theo_Score's picture

Theo_Score | Wed, 02/22/2017 - 00:29

Thank you Daniel,

There are two issues. Firstly, here is the last few lines of my input script;

# Insert the first particles so that dump is not empty
run 1
#dump dmpvtk all custom/vtk 1000 post_0.$a/dump*.data.vtk id type x y z vx vy vz fx fy fz
dump dmp all custom 1000 post_0.$a/dump*.data id type x y z vx vy vz fx fy fz
dump dmpstlcol all mesh/stl 1000 post_0.$a/output*.stl cyl col

# Insert particles
run 70000
fix move all move/mesh mesh cyl linear 0.0 0.150 0
run 20000
unfix move
run 10000

# Post Processing
shell cd post_0.$a
shell python /home/dkc/Installations/LPP/src/lpp.py dump*.data

# Regular loop
next a
clear
jump in.script.coord_number

The script can run and do post-processing but then afterwards, it cant jump to the next stage. How do I end this operation in script so that the loop continues?

Secondly, I have tried to use the dump custom/vtk (inactivated now) as shown on above the dump custom, running the script will flag me that dump custom/vtk is an illegal command. Is there anything wrong with that command or do i need to install additional packages?

Thank you.

Daniel Queteschiner | Wed, 02/22/2017 - 04:49

For dump custom/vtk to work you must install the VTK libraries and development headers and compile LIGGGHTS using the auto Makefile or one of the ubuntuVTK* Makefiles (depending on the VTK version you have installed).

Daniel Queteschiner | Wed, 02/22/2017 - 04:57

My first guess would be that
jump in.script.coord_number
doesn't work because you are in the post_o.$a folder and no longer in the folder of the script, so I would try to add
shell cd ..
to return to the original folder. Could be wrong though as I haven't tried it.

Kathy | Fri, 11/22/2019 - 19:05

Hello All,
I have installed CFDEM coupling and it works perfectly. Using octave we can get all results but there is a problem when the shell reads the following line in the post processing

#- get VTK data from liggghts dump file
cd $casePath/DEM/post
python -i $CFDEM_LPP_DIR/lpp.py dump.liggghts_run

this line will be executed and the conversion to VTK will occurs but the code stalls on this line and does not return to execute the rest of the shell (go to CFD/fomeToVTK folder and open paraview, or etc..)

Here is the output of this line and where everything stops ...

starting LIGGGHTS memory optimized parallel post processing
chunksize: 8 --> 8 files are processed per chunk. If you run out of memory reduce chunksize.
Working with 96 processes...
calculating chunks 1 - 1 of 1
wrote 1 granular snapshots in VTK format
time needed: 0.200386047363 sec
>>>

Any idea why this happens? Any help to debug that will be very much appreciated.