crazy results with vDW df2 [SOLVED]

[recohen]

I am getting crazy results with abinit 8.6.1, with huge negative and positive frequencies fir the acoustic modes, and even the highest frequencies are crazy, like 12000 cm-1 instead of 3000 cm-1. I have relaxed the structure well with abinit and used tight constraints for the SCF WFs. I am using non-local vdW DF2, is it possible that it doesn't work with linear response, although the code runs with no errors? There is no symmetry (P1) and lots of atoms (42).

vdw_xc 2


Phonon frequencies in cm-1 :
- -1.825134E+03 1.782080E+03 5.254005E+03 5.254427E+03 5.254681E+03
- 5.275308E+03 5.296730E+03 5.301214E+03 5.301365E+03 5.313155E+03
- 5.316408E+03 5.318201E+03 5.325414E+03 5.348848E+03 5.351307E+03
- 5.361765E+03 5.383404E+03 5.415318E+03 5.417360E+03 5.418599E+03
- 5.462819E+03 5.504091E+03 5.504399E+03 5.506602E+03 5.507381E+03
- 5.559368E+03 5.585539E+03 5.593731E+03 5.623650E+03 5.689529E+03
- 5.715609E+03 5.720529E+03 5.721033E+03 5.731985E+03 5.739028E+03
- 5.741687E+03 5.743078E+03 5.777978E+03 5.832430E+03 5.836022E+03
- 5.836689E+03 5.837354E+03 5.889213E+03 5.912300E+03 5.914306E+03
- 5.915372E+03 5.915953E+03 5.922249E+03 5.948173E+03 6.054346E+03
- 6.057390E+03 6.058969E+03 6.093693E+03 6.126141E+03 6.150058E+03
- 6.161745E+03 6.165646E+03 6.167311E+03 6.226902E+03 6.238943E+03
- 6.245186E+03 6.248195E+03 6.248904E+03 6.249851E+03 6.299552E+03
- 6.303278E+03 6.341647E+03 6.352635E+03 6.369772E+03 6.404858E+03
- 6.504701E+03 6.535371E+03 6.617583E+03 6.619535E+03 6.621239E+03
- 6.623210E+03 6.640542E+03 6.689423E+03 6.695001E+03 6.753421E+03
- 6.788170E+03 6.832578E+03 6.833641E+03 6.851318E+03 6.906809E+03
- 6.939565E+03 6.940615E+03 6.955927E+03 7.107984E+03 7.130676E+03
- 1.109201E+04 1.109400E+04 1.110244E+04 1.111064E+04 1.112498E+04
- 1.112752E+04 1.113922E+04 1.116887E+04 1.117761E+04 1.119263E+04
- 1.160630E+04 1.161034E+04 1.161790E+04 1.162015E+04 1.162241E+04
- 1.162493E+04 1.162875E+04 1.164506E+04 1.164512E+04 1.164617E+04
- 1.165481E+04 1.165612E+04 1.165955E+04 1.167740E+04 1.168984E+04
- 1.169311E+04 1.170878E+04 1.172319E+04 1.172477E+04 1.173000E+04
- 1.203445E+04 1.203556E+04 1.208933E+04 1.209395E+04 1.212206E+04
- 1.212912E+04

[recohen]

Well I ran without the vdw flag, and get the same results, so it seems that vdw_xc 2 was just ignored with no error. No wonder results are terrible.

[mverstra]

Hello,

indeed, the phonons are coded with the Grimme D3 functional (and work fine) but not the DF2, which is basically in production for the ground state, but has not yet been added to the DFPT. Do you need the DF in particular or just a decent vdw (ie D3 is enough)?

Your phonons are probably bad because the GS and DFPT energies are inconsistent, the former including DF2. I'll see about adding a warning.