QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: beauyy on December 13, 2022, 13:32

Title: Difficult to converge in spin-polarized two-probe system
Post by: beauyy on December 13, 2022, 13:32

Hi all,

I am calculating the zero-bias transmission spectra of two-probe MBHT system. It is very difficult to converge and very low-efficient, even though I have set many computational parameters higher than the guide in https://docs.quantumatk.com/manual/technicalnotes/spin_polarized_negf_convergence_guide/spin_polarized_negf_convergence_guide.html.  To make it converge more quick, I have even reduced the convergence criteria, but still failed to accelerate the calculation. The attachment is my script. 
Is there any method for me to achieve a faster and more reliable convergence? It seems that SGGA  calculation spends too much time. If I choose LSDA, is there any method to achiev a faster and more reliable convergence.