Sir,
I am trying to do charge equilibration using fix qeq/reax command of lammps. For that, i modeled activated epon with both ends hydrogen is removed and packed in 200x200x200 Angstrom box with PCFF force field (Data file is attached with this mail).
The modeling of epon is done using material studio 8.0 with only geometry optimized using PCFF force field. The model is then exported in the form of car and mdf files. The data file is generated using msi2lmp by using the following command line:
msi2lmp.exe packactepon -p 0 -c 2 -frc pcff.frc -i -shift 0.0 0.0 0.0
From this command, the data file is generated as attached herewith.
Now, for charge equilibration I use the input script as defined in the example directory of lammps (input script attached with this mail).
I run this script on np 4 and i get the error message as:
job aborted:
rank; node: exit code[: error message]
0: HostName: -1073741819: process 0 exited without calling finalize
1: HostName: 123
2: HostName: -1073741819: process 0 exited without calling finalize
3: HostName: -1073741819: process 0 exited without calling finalize
I am totally confused whats the error is?
I also tried to look this type of error on previous mails to lammps-user but didn’t find anything useful.
I also searched for google, again same result, didn’t find anything useful.
Any information and help will be a great favor for me. Please help me with this.
Thanking you.
1.in (1.99 KB)
log.cite (454 Bytes)
log.lammps (3.66 KB)
packactepon.data (28.9 KB)
Sir,
I am trying to do charge equilibration using fix qeq/reax command of lammps. For that, i modeled activated epon with both ends hydrogen is removed and packed in 200x200x200 Angstrom box with PCFF force field (Data file is attached with this mail).
The modeling of epon is done using material studio 8.0 with only geometry optimized using PCFF force field. The model is then exported in the form of car and mdf files. The data file is generated using msi2lmp by using the following command line:
msi2lmp.exe packactepon -p 0 -c 2 -frc pcff.frc -i -shift 0.0 0.0 0.0
From this command, the data file is generated as attached herewith.
Now, for charge equilibration I use the input script as defined in the example directory of lammps (input script attached with this mail).
I run this script on np 4 and i get the error message as:
job aborted:
rank; node: exit code[: error message]
0: HostName: -1073741819: process 0 exited without calling finalize
1: HostName: 123
2: HostName: -1073741819: process 0 exited without calling finalize
3: HostName: -1073741819: process 0 exited without calling finalize
I am totally confused whats the error is?
this is just from the MPI library telling you that your calculation
stopped unexpectedly.
I also tried to look this type of error on previous mails to lammps-user but didn't find anything useful.
I also searched for google, again same result, didn't find anything useful.
Any information and help will be a great favor for me. Please help me with this.
try running in serial and watch the output on the screen; you should
get an error message telling your what your exact error is.
axel.
I tried to run it in serial. There is neither a warning nor a error message. But, now i am not getting the previous error message.
Now, the lammps stopped working. and closes automatically.
I am attaching the previous files again with the param.qeq.reax file which i forgot to upload with my previous mail.
Thanking you.
Sumit kumar Sinha
Research Scholar
Malaviya National Institute of Technology JaipurJaipur, Rajasthan-302017
India
param.qeq.reax (235 Bytes)
1.in (1.99 KB)
packactepon.data (28.9 KB)
log.cite (454 Bytes)
log.lammps (3.65 KB)
I tried to run it in serial. There is neither a warning nor a error message. But, now i am not getting the previous error message.
Now, the lammps stopped working. and closes automatically.
I am attaching the previous files again with the param.qeq.reax file which i forgot to upload with my previous mail.
ok. that was causing the error, i saw previously. which is now
resolved and instead i see a segmentation fault due to out-of-range
array access.
please try using fix qeq/shielded instead of fix qeq/reax. it
implements the same algorithm, but is a cleaner/newer implementation.
it looks like something in your input triggers a bug in fix qeq/reax
when used without reax/c.
the parameter file for fix qeq/shielded uses the same parameters as
fix qeq/reax, but you need to add two more columns that will be
ignored with fix qeq/shielded, but only used by other fix qeq
variants. for details please review the documentation of fix
qeq/shielded.
axel.
Hi,
Does it means that it is preferable to use fix qeq/shielded than qeq/reax ? I already got segmentation fault with real/c calculation. I will check deeper.
But I think that in any case reax/c ask for qeq/reax ?If yes how to avoid this using qeq/shielded ?
Bes retards
Pascal
Sir,
Thanks a lot. It works with qeq/shielded. I just need to add two more columns to my parameter file.
I just don’t know how to thanks axel sir. He is great. Solved my problem in just couple of minutes which I am trying to do from last 3 days.
Sorry, if I spam the lammps user mail list. But, I just want to thank him. Please forgive me.
Thanking you.
Sumit Kumar Sinha
Research Scholar
Malaviya National Institute of Technology Jaipur
Jaipur, Rajasthan-302017
India
Hi,
Does it means that it is preferable to use fix qeq/shielded than qeq/reax ?
this is a question for the author of the QEQ package, ray shan (cc'd).
the documentation suggests, that this transition is intended. to make
it functionally equivalent, i.e. allow to extract the qeq parameters
directly from pair style reax/c and to circumvent the check for fix
qeq/reax in pair style reax/c some additional programming is needed.
this is not too complicated and in my superficial testing, it seems to
be working well enough.
so i've submitted a pull request with the required changes on github at:
https://github.com/lammps/lammps/pull/1310
I already got segmentation fault with real/c calculation. I will check deeper.
But I think that in any case reax/c ask for qeq/reax ?If yes how to avoid this using qeq/shielded ?
with a computer, a screen, a keyboard, a compiler, and a good text editor. 
Bes retards
retards? auto-correct malfunction??
ciao,
axel.
we've tagged LAMMPS version 1 February 2019 a few minutes ago and that
version includes changes so fix qeq/shielded can be used instead of
fix qeq/reax. the command line of the two fixes is slightly different,
but fix qeq/shielded can now also directly read the qeq parameters
from pair style reax/c. examples for using both fixes are in
examples/reax/
axel.
Hi Axel, Ray
Thanks a lot
Pascal
we've tagged LAMMPS version 1 February 2019 a few minutes ago and that
version includes changes so fix qeq/shielded can be used instead of
fix qeq/reax. the command line of the two fixes is slightly different,
but fix qeq/shielded can now also directly read the qeq parameters
from pair style reax/c. examples for using both fixes are in
examples/reax/
axel.