<br clear="all">Dear Prof. P. Blaha and all Wien2k users,<br><br>I'm student of physics at the Federal University of Sao Joao del Rei where I work in Group of New Materials.<br>For my calculations, I used a copy of Wien2k_09.2 compiled under a 64bit architecture with the Intel compiler.<br>
My studies are based on defects in the crystal FeSi2 in its beta phase. My unit cell (case.struct) has 48 atoms, impurity and perfect crystal, and 47 atoms, vacancies.<br>My purpose is to relax when the first neighbors around the defects that I tease, impurity and vacancy.<br>
For that I have based on the example given in the manual Wien2k 10.3, page 170.<br>Before starting my calculations, I made the examples given in Chapter 10 and got success in your results.<br>Soon after I began my crystal betaFeSi2. With case.struct below I boot the LAPW calculation (init_lapw) with 3% reduction of the RMT using the GGA with an energy -7.0Ry and 100k-points. After this initialize, the calculation auto squeegee concistente (SCF, run_lapw fc -1.0 -it) with convergence in the force 1mRy.<br>
Initiate the calculation, and the cycles being made (dstart, lapw0, lapw1, lapw2, core, mixer) my case.vector in my Scratch, increases until it reaches an absurd size of 15GB, where the SCF is stopped and the account for.<br>
<br>Could someone help me with this problem. If you need more details just contact.<br>-- <br><br>Kind regards,<br>_____<br>Ranyele Amorim Martins<br>Grupo de Novos Materiais<br>Univ. Federal de São João del Rei - UFSJ<br>