Dear pw_forum I try to simulate a metallic slab under a electric field. In order to improve the convergence, I checked this web page :
http://blog.sina.com.cn/s/blog_5f15ead20100cw4d.html I tried all the suggestion (except band number increase) and the process continue with a very slow convergence, after 283 iterations the ethr=1.2E-9. The important part of the input file is : &CONTROL calculation = 'scf' ... tefield = .true. dipfield = .true. / &SYSTEM ibrav = 0 celldm(1) = 1.0 nat = 21 ntyp = 1 ecutwfc = 30 ecutrho = 500 degauss = 0.03 occupations = 'smearing' smearing = 'mv' nspin = 2 starting_magnetization= 0.1 edir = 3 emaxpos = 0.9 eopreg = 0.2 eamp = 0.019447 / &ELECTRONS electron_maxstep = 300 mixing_mode = 'local-TF' mixing_beta = 0.1 mixing_ndim = 10 / ... K_POINTS (automatic) 6 6 6 0 0 0 ... This process use approx 5Gb of data in the associated files. My question are, -If I increase the band number, The space disk will be more? -Do you have suggestion about how to proceed in order to avoid this I/O increase. and improve the convergence or velocity. -I'm (over/under)estimating some previous value?, k-points, mixing_*, ecut*? Thanks in advance Joaquin Peralta Materials Science and Engineering Iowa State University. -- ---------------------------------------------------- Group of NanoMaterials ---------------------------------------------------- http://www.gnm.cl ---------------------------------------------------- Joaqu?n Andr?s Peralta Camposano ---------------------------------------------------- http://www.lpmd.cl/jperalta In a world without frontiers, who needs Gates and Win. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20100827/16bd0a62/attachment.htm
