On Mon, Jun 18, 2018 at 11:50 AM, SUBSCRIBE AMP-USERS Mingxia Zhou
<[log in to unmask]> wrote:
> Hi everyone,
> My name is mingxia, I am a new user for Amp. However, I had some errors when install the Amp.
> Follow the installation instruction online, I think I have set the environment successfully. The problem occurs in building fortran modules. And when I run the tests, the nosetests failed. After typing f2py -c -m fmodules model.f90 descriptor/cutoffs.f90 descriptor/gaussian.f90 descriptor/zernike.f90 model/neuralnetwork.f90, the output message is shown as:
> f2py -c -m fmodules model.f90 descriptor/cutoffs.f90 descriptor/gaussian.f90 descriptor/zernike.f90 model/neuralnetwork.f90
> running build
> running config_cc
> unifing config_cc, config, build_clib, build_ext, build commands --compiler options
> running config_fc
> unifing config_fc, config, build_clib, build_ext, build commands --fcompiler options
> running build_src
> copying /soft/pymatgen/180120/lib/python3.6/site-packages/numpy/f2py/src/fortranobject.c -> /tmp/tmpsjmi4xnt/src.linux-x86_64-3.6
> copying /soft/pymatgen/180120/lib/python3.6/site-packages/numpy/f2py/src/fortranobject.h -> /tmp/tmpsjmi4xnt/src.linux-x86_64-3.6
Here the python version being used for compiling the Fortran module is 3.6.
> Warning: Unused module variable 'no_nodes_of_elements' which has been explicitly imported at (1)
> /usr/bin/gfortran -Wall -g -Wall -g -shared /tmp/tmpsjmi4xnt/tmp/tmpsjmi4xnt/src.linux-x86_64-3.6/fmodulesmodule.o /tmp/tmpsjmi4xnt/tmp/tmpsjmi4xnt/src.linux-x86_64-3.6/fortranobject.o /tmp/tmpsjmi4xnt/model.o /tmp/tmpsjmi4xnt/descriptor/cutoffs.o /tmp/tmpsjmi4xnt/model/neuralnetwork.o /tmp/tmpsjmi4xnt/descriptor/gaussian.o /tmp/tmpsjmi4xnt/descriptor/zernike.o /tmp/tmpsjmi4xnt/tmp/tmpsjmi4xnt/src.linux-x86_64-3.6/fmodules-f2pywrappers2.o -lgfortran -o ./fmodules.cpython-36m-x86_64-linux-gnu.so
> Removing build directory /tmp/tmpsjmi4xnt
Here it seems that the modules were successfully built for _python 3.6_.
> For the nosetests, the error message is shown as:
> /directory of the file//soft/andrewpeterson-amp-a80158ba4871/amp/__init__.py:27: UserWarning: Did not find fortran modules.
> warnings.warn('Did not find fortran modules.')
> Gaussian/Neural non-periodic standard. ... ERROR
> Gaussian/Neural periodic standard. ... ERROR
> Gaussian/tflowNeural non-periodic. ... ok
> Gaussian/tflowNeural periodic. ... ok
> Gaussian/Neural training non-periodic standard test. ... ERROR
> Gaussian/Neural training periodic standard test. ... ERROR
> Guassian/Neural force call. ... ERROR
> Guassian/Neural training. ... ERROR
> Gaussian fingerprints consistency. ... ERROR
> Zernike fingerprints consistency. ... ERROR
> Displaced atom test. ... ok
> FingerprintPlot test. ... ERROR
> Gaussian/Neural numeric-analytic consistency. ... ERROR
> Rotational/translational invariance. ... ok
> /soft/gpaw/Anaconda3-4.2.0-171010/lib/python3.5/site-packages/ase/lattice/surface.py:17: UserWarning: Moved to ase.build
> warnings.warn('Moved to ase.build')
Here, your anaconda is executing scripts with python 3.5!.
> Do you know what's wrong during my installation process and how to solve it? Thank you for your patient and time.
Is there any way you can force your anaconda to use Python3.6 instead
of Python3.5? The problem here is that your Fortran modules are built
for 3.6. I don't use Anaconda myself, but probably this could help
Let us know if that worked.
Muammar W El Khatib Rodriguez
Postdoctoral Research Associate
Brown University School of Engineering
184 Hope Street
Providence, RI, 02912, USA