Bug on one specific case on multi-core

Dear Lammps Users,

I cannot understand why my input file could not be run on multi-core processors while it is working perfectly well in single core. I wonder if there are any commands that I need to add to the input?

The multi-core processors can run my other jobs other than the specific one that I am attaching here, and after testing this specific case on a single core processor, on which it is runnable, so I doubt the problem may lie in this specific case on multi-core processors.

There are no error messages other than the system hang at read data. After ‘number of atoms’ has been read, the system hang at reading ‘number of bonds’…

Please share me with your ideas, as it is running extremely slow on single-core…

Files attached here are: input file and coord. file; an output file from multi-core(output_changebox_hang); a particle log file from single core processor(before6klog.lammps).

I appreciate very much for helping me out here.
Regards,
Ling

coord.melt.changebox.6000 (4.21 MB)

input_meltellip_changebox (904 Bytes)

output_changebox_hang (208 Bytes)

before6klog.lammps.log (2.58 KB)

Dear Lammps Users,

I cannot understand why my input file could not be run on multi-core
processors while it is working perfectly well in single core. I wonder if
there are any commands that I need to add to the input?

The multi-core processors can run my other jobs other than the specific one
that I am attaching here, and after testing this specific case on a single
core processor, on which it is runnable, so I doubt the problem may lie in
this specific case on multi-core processors.

There are no error messages other than the system hang at read data. After
'number of atoms' has been read, the system hang at reading 'number of
bonds'...

sorry, but i cannot reproduce it. this input deck works
perfectly fine for me in parallel using a current version
of lammps.

axel.

Hi Axel,

I am using sep 2012 version, which one are you using to reproduce it?

Thanks very much for your prompt reply!
Ling

Hi Axel,

I am using sep 2012 version, which one are you using to reproduce it?

current always means the version as is available
from the git/svn repository at the moment. so
right now this is version 11 Jan 2013. it also
works find with an older parallel executable
which is version 6 Dec 2012.

Thanks very much for your prompt reply!

thanks for providing a complete(!) input deck.
that makes it easy to quickly double check
a problem.

axel.