[gpaw-users] 7 failed tests

Fri Sep 6 19:40:14 CEST 2013

Hi,

On 09/06/2013 07:14 PM, Rasmus Karlsson wrote:
> (This is a repost from the gpaw-developers list, since I got no 
> replies there)
> Hi!
> After compiling gpaw 0.9.1.10645 with ase version 3.8.0.3237, numpy 
> version 1.3.0 and scipy version 0.7.0 I get a number of errors in the 
> tests. I'm using gpaw-setups v 0.9.9672. The cluster uses nodes with 
> AMD Opteron 6238 (Interlagos) CPUs. Furthermore, I'm using 
> gcc-compiled lapack (3.4.0) and openblas (0.1), and gcc-compiled 
> openmpi v 1.6.4.
>
> Specifically, the following tests failed (error messages further down 
> in the email)
> fileio/hdf5_noncontiguous.py
> pw/mgo_hybrids.py
> lcao_h2o.py
> lcao_force.py
> IP_oxygen.py
> dipole.py
> bse_MoS2_cut.py
>
> I have hdf5 version 1.8.10-patch1, but maybe another version is needed 
> for the latest gpaw. HDF5 is not so important for what I'm doing, so 
> that problem might be acceptable. Anyway, are there any ideas as to 
> why the other 6 tests failed?
>
> These are the errors (relevant parts of the output files) I get for 
> each of the tests:
>
> pw/mgo_hybrids.py
> --------------------
> Traceback (most recent call last):
>   File "pw/mgo_hybrids.py", line 30, in <module>
>     de_skn = vxc(calc, hyb_calc) - vxc(calc, 'LDA')
>   File 
> "/home/r/rasmusk/pfs/programs/gpaw-0.9.1.10645/gpaw/xc/tools.py", line 
> 26, in vxc
>     paw.get_xc_difference(xc)
>   File 
> "/home/r/rasmusk/pfs/programs/gpaw-0.9.1.10645/gpaw/aseinterface.py", 
> line 423, in get_xc_difference
>     return self.hamiltonian.get_xc_difference(xc, self.density) * Hartree
>   File 
> "/home/r/rasmusk/pfs/programs/gpaw-0.9.1.10645/gpaw/hamiltonian.py", 
> line 405, in get_xc_difference
>     xc.calculate_exx()
>   File 
> "/home/r/rasmusk/pfs/programs/gpaw-0.9.1.10645/gpaw/xc/hybridg.py", 
> line 340, in calculate_exx
>     self.calculate_exx_paw_correction()
>   File 
> "/home/r/rasmusk/pfs/programs/gpaw-0.9.1.10645/gpaw/xc/hybridg.py", 
> line 613, in calculate_exx_paw_correction
>     self.evc -= np.dot(D_p, setup.X_p)
> TypeError: unsupported operand type(s) for *: 'float' and 'NoneType'
>
> lcao_h2o.py
> --------------------
> Dipole Moment: [  7.27763290e-14   7.44357242e-14 -4.80857449e-01]
> Energy -10.3905075644
> Ref -10.266984
> Err 0.123523564376
> Traceback (most recent call last):
>   File "lcao_h2o.py", line 24, in <module>
>     assert err < 1e-4
> AssertionError
i'm getting almost the same Err value when using setups 0.8.
Are you sure you use 0.9?
echo $GPAW_SETUP_PATH

Best regards,

Marcin
>
> lcao_force.py
> --------------------
> Forces in eV/Ang:
>   0 O     1.50949    1.69245   -6.66452
>   1 H    -0.91270   -1.53450    3.78840
>   2 H    -0.58858   -0.11066    2.70873
> Force
> [[ 1.50948536  1.69244746 -6.66451859]
>  [-0.91269716 -1.53449586  3.78839681]
>  [-0.58857688 -0.11065799  2.70872717]]
>
> Reference result
> [[ 1.05200088  1.64950565 -5.01149812]
>  [-0.70775684 -0.89551937  3.03218888]
>  [-0.3330413  -0.70650306  1.77179155]]
>
> Error
> [[ 0.45748448  0.04294181  1.65302047]
>  [ 0.20494032  0.63897649  0.75620793]
>  [ 0.25553558  0.59584507  0.93693562]]
>
> Max error
> 1.65302046563
> Traceback (most recent call last):
>   File "lcao_force.py", line 54, in <module>
>     assert err < 3e-3
> AssertionError
>
>
> IP_oxygen.py
> --------------------
> Total Magnetic Moment: 3.000000
> Spin contamination: 0.000000 electrons
> Local Magnetic Moments:
> 0 3.0
>
> 13.965354077
> Traceback (most recent call last):
>   File "IP_oxygen.py", line 23, in <module>
>     equal(e0, -1.88477, energy_tolerance)
>   File 
> "/home/r/rasmusk/pfs/programs/gpaw-0.9.1.10645/gpaw/test/__init__.py", 
> line 25, in equal
>     raise AssertionError(msg)
> AssertionError: -1.69872400425 != -1.88477 (error: |0.186045995755| > 
> 4e-05)
>
> dipole.py
> --------------------
> Dipole Moment: [ -5.90227807e+00   1.28376553e+00 -1.18314288e-08]
> 2.07504659748 2.07692326303
> Ref value of previous calculation 2.07342988218
> Value in this calculation 2.07504659748
> Error 0.00187666554622 0.00161671530144
> Traceback (most recent call last):
>   File "dipole.py", line 120, in <module>
>     assert err2 < 2e-4
> AssertionError
>
> bse_MoS2_cut.py
> --------------------
> Calculating phi_qaGp
>   Finished iq 0 in 0:00:00, estimated 0:00:37 left.
>   Finished iq 16 in 0:00:08, estimated 0:00:29 left.
>   Finished iq 32 in 0:00:15, estimated 0:00:22 left.
>   Finished iq 48 in 0:00:22, estimated 0:00:15 left.
>   Finished iq 64 in 0:00:29, estimated 0:00:08 left.
>   Finished iq 80 in 0:00:36, estimated 0:00:01 left.
> Memory used 248.773438 M
>
> Calculating screening interaction kernel.
>     RPA macroscopic dielectric constant is: 1.008
>   Finished iq 0 in 0:00:14, estimated 0:02:45 left.
>   Finished iq 2 in 0:00:23, estimated 0:02:22 left.
>   Finished iq 4 in 0:00:33, estimated 0:02:13 left.
>   Finished iq 6 in 0:00:42, estimated 0:02:03 left.
>   Finished iq 8 in 0:00:51, estimated 0:01:54 left.
>   Finished iq 10 in 0:01:01, estimated 0:01:44 left.
>
> Calculating BSE matrix elements
>   Finished pair orbital 0 in 0:00:02, estimated 0:02:11 left.
>   Finished pair orbital 16 in 0:00:27, estimated 0:01:44 left.
>   Finished pair orbital 32 in 0:00:53, estimated 0:01:19 left.
>   Finished pair orbital 48 in 0:01:18, estimated 0:00:53 left.
>   Finished pair orbital 64 in 0:01:44, estimated 0:00:27 left.
>   Finished pair orbital 80 in 0:02:08, estimated 0:00:03 left.
> Diagonalizing BSE matrix.
> Use lapack.
> Calculating dielectric function.
>
> Sum rule:
> N1 = 0.736282, -95.909544  % error
> Memory usage: 248.77 MB
>
> ============================================================
> Timing:                               incl.     excl.
> ...
>
> Any suggestions are appreciated.
>
> Rasmus
>

-- 
***********************************

Marcin Dulak
Technical University of Denmark
Department of Physics
Building 307, Room 229
DK-2800 Kongens Lyngby
Denmark
Tel.: (+45) 4525 3157
Fax.: (+45) 4593 2399
email: Marcin.Dulak at fysik.dtu.dk

***********************************