Dear Xavier, Thank you very much for the nice suggestions. I have reduced it by 8% and would see how it works.
Suddhasattwa Ghosh -----Original Message----- From: [email protected] [mailto:wien-bounces at zeus.theochem.tuwien.ac.at] On Behalf Of Rocquefelte Sent: Saturday, March 20, 2010 1:15 PM To: A Mailing list for WIEN2k users Subject: Re: [Wien] [SPAM?] RE: [Wien2k Users] Overlapping sphereserror duringOptimize.job Yes for sure, you should use more than 3% reduction of the RMTs if you plan to consider up to 10% variation of the c/a ratio. However, I would first simply generate the two extrema structures and check the NN distances. Then you simply have to compare the NN distances variation from +10% to 0% to -10% and it will give you a better idea of what you are really doing and which RMT you should use (the smallest one)=
