Dear Nick,
According to your suggestion, I re-read the siesta manual and changed H.tolerance to SCF.H.tolerance, after I re-run the header file, the tolerance is as I set it. Thank you a lot! Best regards?? Lee ------------------ Original ------------------ From: "nickpapior"<nickpap...@gmail.com>; Date: Mon, Oct 15, 2018 09:23 PM To: "siesta-l"<siesta-l@uam.es>; Subject: Re: [SIESTA-L] Quastion about DM.Tolerance and H.Tolerance Dear Lee, Could you show the output? There should be no flag called H.Tolerance, it is named SCF.H.Tolerance. I haven't had any problems with these keys, so my guess is that you have some other flags in your input that overrides, changes the actual convergence criteria. Den man. 8. okt. 2018 kl. 22.08 skrev Patrick Lee <lipeihen...@foxmail.com>: Dear Siesta Users, Thank you for checking this message! I tried to modify DM.Tolerance from 1.d-4 to 1.d-3 and H.Tolerance from 1.d-3 to 1.d-2 in "scat.fdf" when using transiesta to calculate the transmission characteristics of semiconductors. DM.Tolerance 1.d-3 # Tolerance in maximum difference H.Tlerance 1.d-2 After I modified them, in the "filename.out" file, It can be seen that the convergence precision has been modified for the accuracy I set??as shown below.redata: DM tolerance for SCF = 0.001 redata: Hamiltonian tolerance for SCF = 0.010000 eV However, when performing SCF iterations, DM.Tolerance and H.Tolerance still uses the default precision (1.d-4 & 1.d-3) to determine whether to converge. The changes I made does not play a role in the modification of the DM and Hamiltonian tolerance. This means that I still can't reduce the precision of DM and Hamiltonian in order to get a rough result. I don't know how to solve this problem. I will be very grateful if you can help me! Best regards?? Lee -- Kind regards Nick
