QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: davalenciah on November 29, 2011, 14:35

Title: Warning Message
Post by: davalenciah on November 29, 2011, 14:35

Hello

I was doing a DFT transport calculation and I got a message in the output file as:

################################################################################
# WARNING                                                                      #
#                                                                              #
# The computed multigrid residual is greater than the required accuracy.       #
#                                                                              #
# Computed residual :   1.81808e-12                                            #
# Required accuracy :   1.00000e-12                                            #
#                                                                              #
################################################################################

What parameter I should increase to reach the require accuracy?
the mess cut off?
k points or another one?


Thanks for your Help

Title: Re: Warning Message
Post by: Anders Blom on November 29, 2011, 15:49

If the residual is still so small it's absolutely nothing to worry about.

Title: Re: Warning Message
Post by: esp on January 18, 2012, 21:36

what if it is not that small?

The computed multigrid residual is greater than the required accuracy
# Computed residual :   5.79742e-09                                            #
# Required accuracy :   1.00000e-12 

Title: Re: Warning Message
Post by: Anders Blom on January 18, 2012, 21:39

That's still small :)

Title: Re: Warning Message
Post by: Nordland on January 19, 2012, 15:15

It means that another iteration in the multigrid solver would only improve the the solution of the Poisson equation for the electrostatic by a 1e-9 Hartree.

Title: Re: Warning Message
Post by: abhi005 on November 12, 2014, 11:13

my residual error is shown below.what is INF0e.Is this error is quite small or i have to worry?

Computed residual :   1.#INF0e+000                                            #

Title: Re: Warning Message
Post by: Anders Blom on November 12, 2014, 12:58

Infinity is not quite small...

Title: Re: Warning Message
Post by: abhi005 on November 12, 2014, 14:00

thanks anders blom for your reply....actually i didnt know that INF means infinity.