dump molfile outputs: all zeros

LAMMPS LAMMPS Mailing List Mirror
1

Hi community members,

I am trying to use dump molfile to get pdb form of data output. I am getting output but all zeros in following form:

2

Hi community members,

I am trying to use dump molfile to get pdb form of data output. I am getting output but all zeros in following form:

CRYST1 0.000 0.000 0.000 90.00 90.00 90.00 P 1 1
ATOM 1 2 UNK X 0 0.000 0.000 0.000 0.00 0.00 UNK
ATOM 2 2 UNK X 0 0.000 0.000 0.000 0.00 0.00 UNK
ATOM 3 2 UNK X 0 0.000 0.000 0.000 0.00 0.00 UNK
ATOM 4 2 UNK X 0 0.000 0.000 0.000 0.00 0.00 UNK
ATOM 5 2 UNK X 0 0.000 0.000 0.000 0.00 0.00 UNK
.
.
.
ATOM 6249 1 UNK X 0 0.000 0.000 0.000 0.00 0.00 UNK
ATOM 6250 1 UNK X 0 0.000 0.000 0.000 0.00 0.00 UNK
END

I am using following command in input script.

dump DP2 all molfile 2 dump_cpr-*.pdb pdb /Users/…/softwares/vmd-1.9.2/plugins/LINUXAMD64/molfile

I am using lammps version of 9Dec2014 on linux machine. Also copied header files from VMD/…/include/ to USER-MOLEFILE package before making lammps again.

Total ATOM IDs and types are correct though.
Please suggest if I am missing some point while doing all mentioned above.

Before anyone will take a serious look, you have to test with the very latest version of lammps to verify that you are not reporting an issue that has already been addressed.

Axel

3

Hi Axel and others,

I checked it with recent version 15May2015 with VMD 1.9.2. The issue remains same. Though I can get output from usual “dump atom/custom” commands.

Thanks

4

Hi Axel and others,

I checked it with recent version 15May2015 with VMD 1.9.2. The issue remains
same. Though I can get output from usual "dump atom/custom" commands.

i cannot reproduce it with 28 Jul 2015. however, there have not been
any changes to the code since march 2015.

what platform specifically are you running on? is this a standard
compile with one of the standard makefiles (serial or mpi)?

can you try a precompiled serial binary from rpm.lammps.org?
for x86_64 RHEL 6.x for example you would do:

wget http://rpm.lammps.org/centos/6/x86_64/lammps-20150724-12ak.el6.x86_64.rpm
rpm2cpio lammps-20150724-12ak.el6.x86_64.rpm | cpio -ivd
./usr/bin/lmp_g++ -in in.mytestinput

axel.

5

Hi Axel,

I have installed lammps on Helium cluster at university (They have already one installed systemwide but I did it to add some pair potentials and other chunk of codes for specific problem).

It is a centos 5.11 final version x86_64.
I compiled parallel version with openmpi and till now compiled code works fine.

FFT_INC = -DFFT_FFTW3 -I/opt/fftw/3.3.3/include

MPI_INC = -I/opt/openmpi/1.6.3/gcc48/include

As I wanted to use VMD Volmol plugin, I tried to use molfile dump option. I get proper dump by all other options like atom/custom I used so far.

As I told earlier, I am getting dump but with full of zeros except type and ID for all particles.

I also, tried to get same results on my OS X 10.10.3 Yosemite. But it rather gives “no suitable plugin found” though I am giving correct path of molfile directory from my VMDapp in /Applications. The reason might be that I used VMD 1.9.2 dmg which was downloaded from VMD webpage and its 32 bit. To check if this is architecture issue, I shifted to linux cluster about which I mentioned above.

I will now look into your suggestions as well as will try VMD 1.9.2 with Latest lammps source on my most recent ubuntu partition. Will let you know the outcome.

Thanks

6

Hi Axel,

I have installed lammps on Helium cluster at university (They have already
one installed systemwide but I did it to add some pair potentials and other
chunk of codes for specific problem).

It is a centos 5.11 final version x86_64.
I compiled parallel version with openmpi and till now compiled code works
fine.

FFT_INC = -DFFT_FFTW3 -I/opt/fftw/3.3.3/include
MPI_INC = -I/opt/openmpi/1.6.3/gcc48/include

As I wanted to use VMD Volmol plugin, I tried to use molfile dump option. I
get proper dump by all other options like atom/custom I used so far.

that reasoning doesn't make sense to me. VMD reads standard and custom
LAMMPS dumps just fine.
the pdb file format is just plain horrible and loses a lot of
information. if needed, the additional information can be just as well
provided through loading a suitably constructed .psf file.

also, i am not aware of a "volmol" plugin. it doesn't seem to be an
official plugin. why does it need .pdb files?
.pdb is among the worst possible file format for almost all purposes
(and .xyz is almost as bad).

As I told earlier, I am getting dump but with full of zeros except type and
ID for all particles.

well, and i do get coordinates, too. :wink:

I also, tried to get same results on my OS X 10.10.3 Yosemite. But it rather
gives "no suitable plugin found" though I am giving correct path of molfile
directory from my VMDapp in /Applications. The reason might be that I used
VMD 1.9.2 dmg which was downloaded from VMD webpage and its 32 bit. To check
if this is architecture issue, I shifted to linux cluster about which I
mentioned above

since the molfile plugins are compiled binaries, you cannot load
32-bit plugins into a 64-bit binary, you need to compile a 32-bit
binary. also, i usually find it much more convenient to simply copy
the plugin binary (.so file) into the working directory of the LAMMPS
simulation.

axel.