Inconsistency in DFTtoMLFT CrF4 tutorial (Windows 10 vs Linux)

asked by Sergei Butorin (2022/07/31 21:51)

The DFTtoMLFT tutorial (using both output-DFT and script files coming with tutorial) for CrF4 produces quite different results when running it on Windows 10 (64bit) vs Ubuntu 18.04 (64bit) with binaries of the latest version downloaded from the Quanty website. Both the results for the ground state Hamiltonian (eigenvalues) and Cr L2,3 XAS spectra are different. Furthermore, the calculated Cr L2,3 XAS spectra in both cases are also different from the calculated reference spectrum coming with the tutorial. What could be the reason for that? Unfortunately, I have not found the way to attach the figure with the calculated spectra.

Answers

, 2022/08/04 07:54

Hello Sergei,

This is interesting. Sadly the website does not have a way for users to attach figures to forum posts, but could you please post the output of the Eigensystem step from both your Linux and Windows runs? If you feel like it you could also put your figures into something like google drive and share the link here.

When I run the CrF4 example on Windows 10 (64bit) I get 

            <E>       <S^2>     <L^2>     <J^2>     <S_z^3d>  <L_z^3d>  <l.s>     <Neg_Cr>  <Nt2g_Cr> <N_Cr>    <N_L>
     1  -15.06649 2.5447666 11.474229 12.162833 -0.96565  0.7640353 -0.550225 1.351021  1.5939872 8.9450083 15.054992
     2  -15.03943 2.4845871 12.815806 13.248321 -0.900346 0.595546  -0.704964 1.1512497 1.7872458 8.9384955 15.061504
     3  -15.02837 2.6574569 12.790435 15.467202 -0.951847 -0.658837 -0.044114 1.7939243 1.1646722 8.9585965 15.041403

Running the exact same script on Ubuntu (Not sure which version my VM is) I get 

        <E>       <S^2>     <L^2>     <J^2>     <S_z^3d>  <L_z^3d>  <l.s>     <Neg_Cr>  <Nt2g_Cr> <N_Cr>    <N_L>     
  1     29.8242   2.44776   14.0507   15.9821  -0.947421  0.148529 -0.194101  1.14371   1.78998   8.9337    15.0663  
  2     29.9368   2.43656   14.0842   15.8566  -0.009671  0.003526 -0.23709   1.14686   1.78707   8.93393   15.0661  
  3     30.05     2.42379   14.1293   15.9185   0.93265  -0.14866  -0.234368  1.14525   1.78889   8.93414   15.0659

This is certainly odd. The developers go through the forum every few days, so hopefully they will be able to clear this up for us soon.

You could leave a comment if you were logged in.

You are here: Quanty - a quantum many body script language » Forum » data » 2022 » Inconsistency in DFTtoMLFT CrF4 tutorial (Windows 10 vs Linux)

Print/export