multiple file parsing, error with lpp

Submitted by tdl on
ckloss

dear tdl, please have a look at the "postscripts" that come with the standard LIGGGHTS examples. Does the error also occur if you run one of these? If the error persists, please let us know! Thanks

Christoph

Fri, 2 Mar 2012 - 13:02 Permalink
tdl

In reply to by anonymous_stub (not verified)

Hello, I can confirm the same error running liggghts on the "breakage" test, then entering the post directory and running "lpp dump*.break".
Thank you,
Alessandro

Mon, 5 Mar 2012 - 13:40 Permalink
ckloss

In reply to by anonymous_stub (not verified)

Hi Alessandro,

don't know where this comes from - maybe an internal python issue . have you tried the same on another machine?

Christoph

Thu, 8 Mar 2012 - 14:30 Permalink
tdl

In reply to by anonymous_stub (not verified)

Dear Christoph,
no, I did not yet try to replicate it elsewhere. I wanted to test it on my personal pc (running kubuntu) but I was busy and couldn't do it. I will try the next week, and will let you know.
By the way, the error happens while running lpp on a virtual machine, with guest os kubuntu, and using the precompiled liggghts packages from the repository.
Thank you,
Alessandro

Fri, 9 Mar 2012 - 09:59 Permalink

Hello, it's me again with another weird issue.
I have run some test (this time with small dump files).
The problem is, when I run lpp on the data files I
get the following error message

Exception RuntimeError: RuntimeError('cannot join current thread',) in ignored

And the resulting files show a clearly broken
behavior.
The file name themselves appear to be incorrect.
I am simply naming the dump files "step_N" for
a dump at a given time step, and i run lpp on all
the files with "lpp *".
Am I doing something naively wrong?
Thanks
AS

Note: I tried running lpp with --cpunum 1 and various
values of --chunksize: from 1 to 100 (2000 being my
dump files). I never had the problem using a single
dump file. Can it be related to file naming issues?

Note2: Using --chunksize 2000 (as the number of
existing dumps) does not give the error, and the
corresponding files, when seen in paraview, are correct.
I really suppose there is some sort of issues with the
script when multiple files are parsed in different chunks.