GW gives NaN dielectric values

[basant_ali]

When running a G0W0 calculation, the dielectric function is written as NaN instead of numbers as follows:

INCAR:

Code: Select all

LREAL = Auto         
ISTART = 1 
LPLANE = .FALSE.          
ISIF   =  6           
IBRION =  -1           
PREC   =  Normal  
LMAXMIX = 4           
ISMEAR =  0           
SIGMA  =  0.02        
LASPH = True          
GGA=PE 
ALGO = GW0
ENCUT = 400           
NELM = 1            
EDIFF = 1E-8 
NSW    =  0          
NBANDS  = 560 
LSPECTRAL = .True.
NOMEGA    = 50
NBANDSGW= 500
LSORBIT = .TRUE.
LNONCOLLINEAR = TRUE  
SAXIS =  0 0 1 
GGA_COMPAT = False     
NGXF    = 160         
NGYF    = 160         
NGZF    = 280         
LWAVE  = .TRUE.       
LCHARG = .TRUE.       
LORBIT = 11           
NEDOS = 3000  

OUTCAR:

Code: Select all

 HEAD OF MICROSCOPIC DIELECTRIC TENSOR (INDEPENDENT PARTICLE)
 -------------------------------------
 w=     0.000     0.000
            NaN       NaN         NaN       NaN         NaN       NaN
            NaN       NaN         NaN       NaN         NaN       NaN
            NaN       NaN         NaN       NaN         NaN       NaN

        0.000           NaN       NaN dielectric  constant
 
 w=     1.843     0.000
            NaN       NaN         NaN       NaN         NaN       NaN
            NaN       NaN         NaN       NaN         NaN       NaN
            NaN       NaN         NaN       NaN         NaN       NaN

        1.843           NaN       NaN dielectric  constant
 
 w=     3.680     0.000
            NaN       NaN         NaN       NaN         NaN       NaN
            NaN       NaN         NaN       NaN         NaN       NaN
            NaN       NaN         NaN       NaN         NaN       NaN

        3.680           NaN       NaN dielectric  constant

[manuel_engel1]

Hi,

Thanks for reaching out. This is certainly strange. I will try to investigate what is going on in your calculation. In the meantime, could you tell me if you have run other G0W0 calculations and whether they also produce these NaN results?

[basant_ali]

Thank you for your reply.
I have only tried the same job with lower and higher KPOINTS, all gave the same error.

[basant_ali]

Also,
This GW job is based on a WAVEDER from an Optics job with Algo=Normal or Algo=D instead of Algo=Exact since using Algo=Exact gives an error
scaLAPACK: Routine ZPOTRF ZTRTRI failed! 3841 2 1
LAPACK: Routine ZTRTRI failed!

I am not sure if this might be related to the Nan error.

Thank you,
Basant