Hello, I am running a LDA single point calculation using SIESTA for a system with 1561 atoms.
I selected OrderN as solution method and set "MD.NumCGsteps 0". >From the log, I got the following: *************************************** ...... cgwf: iter = 1000 grad = -0.013328 Eb(Ry) = -4002.562525 cgwf: Maximum number of CG iterations reached ..... siesta: iscf = 1 And the calculation continues but the energy blowup: ordern: enum = 7456.0000 cgwf: iter = 1 grad = -721.775953 Eb(Ry) = -2943.046646 cgwf: iter = 2 grad = -1932.876535 Eb(Ry) = -3016.685582 cgwf: iter = 3 grad = -1587.688987 Eb(Ry) = -3040.309435 cgwf: iter = 4 grad = -803761617.222101 Eb(Ry) = -7438.384633 cgwf: iter = 5 grad = -736299206.349643 Eb(Ry) = -7446.898274 cgwf: iter = 6 grad = ****************** Eb(Ry) = ************** cgwf: WARNING: E. increased; resetting CG cgwf: iter = 7 grad = ****************** Eb(Ry) = ************** cgwf: WARNING: E. increased; resetting CG cgwf: iter = 8 grad = ****************** Eb(Ry) = ************** cgwf: WARNING: E. increased; resetting CG cgwf: iter = 9 grad = -Infinity Eb(Ry) = ************** cgwf: iter = 10 grad = NaN Eb(Ry) = NaN cgwf: iter = 11 grad = NaN Eb(Ry) = NaN cgwf: iter = 12 grad = NaN Eb(Ry) = NaN cgwf: iter = 13 grad = NaN Eb(Ry) = NaN cgwf: iter = 14 grad = NaN Eb(Ry) = NaN cgwf: iter = 15 grad = NaN Eb(Ry) = NaN cgwf: iter = 16 grad = NaN Eb(Ry) = NaN .... *************************************** My doubts: - how to do a single point calculation with OrderN? - Should I use MaxSCFIterations and ON.MaxNumIter at the same time? []'s, Camps
