I am trying to calculate anisotropic super. gap for NbSe2 using epw code. After successful wannierization and calculation of el. spectral function program attempts to calculate imaginary ale real part of gap for set of temperatures. However, it fails to calculate the real part end returns just NaN and also file of NaN-s for the real part in Pade approximation. Then it crashes without any error explanation. I have tried to use different values of npade (0.9 and 0.3) but without any success.
I attach input and output of epw.
Any ideas how to fix this are welcomed
Size of allocated memory per pool: ~= 0.3046 Gb
iter ethr znormi deltai [meV]
1 1.628741E+00 1.723843E+00 4.850853E-01
2 2.236997E+01 1.725398E+00 8.923527E-01
3 9.328945E-01 1.724566E+00 7.818203E-01
4 7.088212E-02 1.724056E+00 8.099794E-01
5 2.074043E-02 1.723880E+00 8.273002E-01
6 1.774101E-02 1.724046E+00 8.163017E-01
7 1.909708E-02 1.723768E+00 8.406679E-01
8 5.528492E-03 1.723782E+00 8.410110E-01
9 7.080690E-03 1.723771E+00 8.413647E-01
10 2.438119E-03 1.723773E+00 8.408701E-01
Min. value of superconducting gap = 0.000000 meV
Convergence was reached in nsiter = 10
iaxis_imag : 36126.67s CPU 36433.56s WALL ( 1 calls)
Pade approximant of anisotropic Eliashberg equations from imaginary-axis to real-axis
Cutoff frequency wscut = 1.0000
pade Re[znorm] Re[delta] [meV]
5540 NaN NaN
Convergence was reached for N = 5540 Pade approximants
Min. value of superconducting gap = 0.000000 meV