UT theoretical chemistry code forum

Posted: **Sun Aug 09, 2015 4:17 pm**

Hi,

I followed the instructions on your website. Made STUBS/libmpi_stubs.a and liblammps_serial.a using the 15May15 version of LAMMPS, copied them in, then compiled eonclient with `make LAMMPS_POT=1`. At the end of the compilation, it printed "POTENTIALS: [...] +LAMMPS [...]", yet when I go to the `eon/client/test/lammps` directory and run eonclient, I get this:

Code: Select all

EON Client
VERSION: r2282
BUILD DATE: Sun Aug 9 16:49:55 BST 2015

Hostname: ...
OS: Linux
Arch: x86_64
PID: 4129
DIR: /home/matt/eon/client/tests/lammps

Loading parameter file config.ini
* [Main] job: minimization
* [Potential] potential: lammps
* [Optimizer] opt_method: cg
* [Optimizer] converged_force: 1e-8
* [Optimizer] lbfgs_inverse_curvature: 0.04

Beginning minimization of pos.con
[Matter]       Iter       Step size           ||Force||         Energy
Unknown Potential: lammps

Running `eon` in the `akmc-fe-lammps` example just gives this before exiting:

Code: Select all

Eon version svn revision 2282
Registering results
Processed 0 results
Queue contains 0 searches
Making 1 process searches
job failed: .//jobs/scratch/0_4:
Created 1 searches
Currently in state 0 with confidence 0.000000

Any ideas?

Many thanks,

Matt

Posted: **Mon Aug 10, 2015 8:05 pm**

I followed the instruction on this page and it works now:
http://theory.cm.utexas.edu/eon/lammps_ ... -potential

The command to compile lammps library was updated:
make serial mode=lib

Posted: **Tue Aug 11, 2015 12:06 pm**

Thanks. I didn't have any problem compiling liblammps_serial.a though (I tried renaming it to liblmp_serial.a just in case, because that's what the EON page calls it, but then eonclient doesn't compile).

Since you've got it working though, I'll delete everything and start again and hopefully figure it out...

Posted: **Tue Aug 11, 2015 12:22 pm**

Excellent! It worked second time. Sorry to have bothered you about it. I can only speculate that I initially made without LAMMPS support, and maybe some files from that previous compilation were left over, even though it reported compiling +LAMMPS at the end, because this second time I tried to compile with LAMMPS support (after deleting+re-dowloading EON) it took a lot longer. Just thought I'd leave the explanation here in case it helps someone else who had the same problem I did.

Edit: Also, eon was giving an "ImportError: No module named mcamc" error. I'd used "python setup.py install" but it hadn't copied the "mcamc" or "yaml" folders into my "/usr/local/lib/python2.7/dist-packages/eon-0.0.0-py2.7.egg/eon/" folder. I copied them in manually and it seems to work fine now, so I'm not sure why that happened. I'm on standard 14.04 Ubuntu. It's all working fine for me now, just mentioning it in case it's an actual bug.

Posted: **Tue Aug 11, 2015 1:47 pm**

Ok great, we'll check the install script.

UT theoretical chemistry code forum

"Unknown Potential: lammps" even though eonclient built with LAMMPS support

"Unknown Potential: lammps" even though eonclient built with LAMMPS support

Re: "Unknown Potential: lammps" even though eonclient built with LAMMPS support

Re: "Unknown Potential: lammps" even though eonclient built with LAMMPS support

Re: "Unknown Potential: lammps" even though eonclient built with LAMMPS support

Re: "Unknown Potential: lammps" even though eonclient built with LAMMPS support