[lammps-users] bug with "centro" dumps

LAMMPS LAMMPS Mailing List Mirror
1

Hi all.

I have a problem which I've checked against the 10Nov05 release (maybe someone could check it on a newer build if they have one handy?)

I get a warning followed by a segfault if I make more than one dump custom with a centro attribute.

2

Craig,

Why not undump the first centro dump after the first
run to avoid having two of them dumping
simultaneously? This avoids the warning message and
the crash. (See modified input script below.)

Paul

# 3d Lennard-Jones melt

units lj
atom_style atomic

# Note that unit density fcc has nearest neighbors at
2^(1/6)
# which is the LJ zero force configuration
lattice fcc 1.0
region box block 0 8 0 8 0 4
create_box 1 box
create_atoms 1
mass 1 1.0

velocity all create 10.0 87287

pair_style lj/cut 1.5
pair_coeff 1 1 1.0 1.0 1.5

neighbor 0.3 bin
neigh_modify every 20 delay 0 check no

fix 1 all nvt 10.0 10.0 100.0

dump 1 all atom 5000 dump.melt.*
dump centroMelt all custom 1000
dump.centro.* tag centro

timestep .0025
thermo 500
run 5000

undump 1
undump centroMelt

unfix 1
fix 2 all nve
fix 3 all viscous 1.0

thermo 500
dump 2 all atom 5000 dump.melt.*
dump centroQuench all custom 1000
dump.centro.* tag centro

run 5000

Hi all.

I have a problem which I've checked against the
10Nov05 release
(maybe someone could check it on a newer build if
they have one handy?)

I get a warning followed by a segfault if I make
more than one dump
custom with a centro attribute.

---------------------
WARNING: More than one dump custom with a centro
attribute
--------------------

I have two separate runs in the input file which is
why I need to
issue the "dump custom centro" command twice.

Wondering if someone else could verify?

Might take a look at the lammps src when I get a
chance.

Thanks,
Craig

input script follows:
---------------------------
# 3d Lennard-Jones melt

units lj
atom_style atomic

# Note that unit density fcc has nearest neighbors
at 2^(1/6)
# which is the LJ zero force configuration
lattice fcc 1.0
region box block 0 8 0 8 0 4
create_box 1 box
create_atoms 1
mass 1 1.0

velocity all create 10.0 87287

pair_style lj/cut 1.5
pair_coeff 1 1 1.0 1.0 1.5

neighbor 0.3 bin
neigh_modify every 20 delay 0 check no

fix 1 all nvt 10.0 10.0 100.0

dump 1 all atom 5000 dump.melt.*
dump centroMelt all custom 1000
dump.centro.* tag centro

timestep .0025
thermo 500
run 5000

unfix 1
fix 2 all nve
fix 3 all viscous 1.0

thermo 500
dump 2 all atom 5000 dump.melt.*
dump centroQuench all custom 1000
dump.centro.* tag centro

run 5000

Using Tomcat but need to do more? Need to support
web services, security?
Get stuff done quickly with pre-integrated
technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1
based on Apache Geronimo

http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642

_______________________________________________
lammps-users mailing list
[email protected]

https://lists.sourceforge.net/lists/listinfo/lammps-users

3

Great, thanks.

Didn't realize one could "undump". It's a bit counter-intuitive, since one needs to explicitly *add* another dump to get output during the second run.

--Craig

4

Actually, I believe that all output commands carry over from run to run.
So every subsequent run in your input file below should write to
the same dump file (or at least maintain the same options, since it looks
like you are dumping to multiple files)

-N

5

Actually, it looks like there is a segfault which comes even if the second centro dump is turned off.

Modified script:

6

Craig,

I ran this script successfully here --- no segfault.
Could you try again with the latest fully patched
version?
http://www.cs.sandia.gov/~sjplimp/download.html
(Download the "LAMMPS upgrade" version.)

Paul

7

Re: LAMMPS crashing when centro was being
printed by dump custom and an unfix was done between
2 runs ... this was a bug in dump_custom.cpp. It will
be fixed in the next release.

Steve