possible bug in fix_srd.cpp

Please post to the list, not to me.

I have a total of 471 big particles, so I expect the above line to be
printed to a maximum of 471 times (1 particle per bin) at any given step.
But I got the above line printed several thousands of time ..

No, you will get that line once for each time each particle overlaps
a bin. If one particle overlaps 20 bins, then you will get the message
20*N times.

I want a script that generates the error that thee are too many big
particles in a bin. Does that one do it?

Steve

Hi Steve,

yes, this script generates the error message - too many big particles in a bin.

I have attached the files and log file containing the error at the end of file compiled with the latest 22Mar2013 version.

Thank you very much.

in.srd_enm (1.87 KB)

input.dat (141 KB)

slurm-205424.out (85.1 KB)

ok - will be next week before I can look at this

Steve

There are no interactions between big/big particles in your
script. What will keep them from collapsing on top of each
other, leading to the error message?

Did you visualize the snapshots you output? Presumably
you would see this collapse.

Steve

Hi Steve,

In my input file I have defined the bonds, bond distances and the bond coefficients between big/big particles in the molecule.
My script file has the line - bond_style harmonic - so the interactions between big/big particles are calculated.
I have visualize the system and the big molecule seems to be doing fine.

Thanks.

ok - didn’t see the bond harmonic command.

I don’t really see how defining a few bonds between 500 monomers
will keep it from overlapping with itself. If it were a flexible
linear polymer chain, it wouldn’t prevent a collapse.

Your script takes far too long for me to run and debug.
I suggest you run to nearly the point when you get the
error message, write a restart file, convert it to a data
file (or use the new write_data command), and run again
to see if it still fails.

Ideally, I want a problem I can run on 1 (or very few procs) in
about 10 seconds to see the problem.

Steve