Dear all,
I am running my case on multi-processors. I want to visualize my results using vmd. But the results are not organized in an ordered way. So I need to sort the results. However, after I sorted it, vmd can not recognize the format anymore.
Anybody can give me some suggestion?
Thanks
Ziyuan
Dear all,
I am running my case on multi-processors. I want to visualize my results
using vmd. But the results are not organized in an ordered way. So I need to
sort the results. However, after I sorted it, vmd can not recognize the
format anymore.
Anybody can give me some suggestion?
please explain your problem in a way that somebody
else than you can understand what is going on.
what "results" are you talking about? i can only guess
that you are referring to trajectory files.
it is not very clear to why you need to sort the trajectories.
cheers,
axel.
Hi, Axel
Sorry for the confusion. I usually write my input script like this:
dump 1 atom 1000 dump.mywater
Then I wrote script as the tools (lmp2vmd) menu suggested on vmd to visualize how my water moves in the simulation.
My problem is the my dump.mywater file is not arranged in an ordered way due to the parralle simulation. So vmd can not read the file. So I sorted my file, but vmd still can’t recognize.
Thanks for the help.
Ziyuan
Hi, Axel
Sorry for the confusion. I usually write my input script like this:
dump 1 atom 1000 dump.mywater
ok.
Then I wrote script as the tools (lmp2vmd) menu suggested on vmd to
visualize how my water moves in the simulation.
there are no scripts in lmp2vmd, only are README.txt
file pointing you towards using topotools.
My problem is the my dump.mywater file is not arranged in an ordered
way due to the parralle simulation. So vmd can not read the file. So I
sorted my file, but vmd still can't recognize.
that is nonsense. vmd ran read your dump file just fine.
there is no need at all to sort it. the lammps trajecory
reader will sort the atoms for you upon reading.
did you tell VMD that you have a lammps trajectory dump?
because, if it is not called something.lammpstrj VMD will
not know that it is a lammps dump and fall back to trying
to parse it as a .pdb file, which will of course fail.
cheers,
axel.
p.s. this kind of question should go to the VMD mailing list, btw.
Hi Axel
Thanks for your answer. What I meant script, I meant readme.txt. I wrote script on vmd according to readme.txt. Sorry that I didn’t make my question clear. VMD can read the dump file fine, but the visualization is messed up. What I meant is that if I get my results from single processor simulation, I have no problem to visualize the water, bonds and so on. But if I run it on multi-processors, after VMD read the dump file, it gave a messed visualization. That’s why I think I need to sort the results in an ordered atom number.
Thanks
ziyuan
Hi Axel
Thanks for your answer. What I meant script, I meant readme.txt. I wrote
script on vmd according to readme.txt. Sorry that I didn't make my question
clear. VMD can read the dump file fine, but the visualization is messed up.
What I meant is that if I get my results from single processor simulation,
I have no problem to visualize the water, bonds and so on. But if I run
it on multi-processors, after VMD read the dump file, it gave a messed
visualization. That's why I think I need to sort the results in an ordered
atom number.
i think i understand what your issue is.
try reading the dump file directly without the .psf. is it still mixed up?
it would probably be much easier for you (and faster, too)
if you would write your trajectory as a .dcd format file.
what version of VMD are you using?
cheers,
axel.
Hi, Axel
My VMD version is 1.8.7 (August 1st, 2009).
I did output dcd format, but vmd can’t load this file.
It seems that the files VMD can load are atom and xyz format.
I read dump file directly without the .psf, it is still mixed up.
If I run on single processor, the visualization is OK, but it’s too slow to get the results.
Thanks
Ziyuan
Hi, Axel
My VMD version is 1.8.7 (August 1st, 2009).
I did output dcd format, but vmd can't load this file.
that is not correct. vmd reads dcd files produced with
LAMMPS just fine. i use them all the time. please provide
proper proof that something is broken.
It seems that the files VMD can load are atom and xyz format.
it can also read .xtc and custom dumps.
I read dump file directly without the .psf, it is still mixed up.
If I run on single processor, the visualization is OK, but it's too
slow to get the results.
that doesn't make much sense.
please provide a way to reproduce this.
a.
Dear all,
I read dump file directly without the .psf, it is still mixed up.
I guess the problem with xyz dump files is that, quoting lammps
documentation: "When LAMMPS is running in parallel, the atom
information written to dump files may be written in an indeterminate
order. It is the case for the xyz style *if the dump group is not
all*."
Indeed, I checked that when trying to visualize a xyz dump file for a
subset of atoms, produced in parallel, VMD only reads atom types for
the first snapshot where the atoms are in a given order, and then
assume that the order is conserved. As it is not the case, atoms seems
to change type during the movie. One solution is to use group all, or
indeed to go to other dump styles.
If I run on single processor, the visualization is OK, but it's too
slow to get the results.
I think Ziyuan simply means lammps computing is slower when run in serial...
Best,
Laurent
Dear all,
>> I read dump file directly without the .psf, it is still mixed up.
I guess the problem with xyz dump files is that, quoting lammps
documentation: "When LAMMPS is running in parallel, the atom
information written to dump files may be written in an indeterminate
order. It is the case for the xyz style *if the dump group is not
all*."
this was about native atom style dumps, not xyz format.
a.