[Wien] LAPW2: semicore band-ranges too large

2015-02-15 Thread delamora
Dear WIEN2k colleagues, The ghost bands are a problem, but I have not found how to get rid of them. One simple example, cubic BaRuO3 Running with runsp -so -orb When r*k=7 or 8 (k-points=100) it finishes without problem but when it is increased to r*k=9

Re: [Wien] LAPW2: semicore band-ranges too large

2015-02-15 Thread Peter Blaha
You have very different sphere radii. In particular, the Sr sphere is too big when you want to use a big RKmax. When you reduce this sphere to eg. 2.2, you can also run RKmax=9. PS: RMTs and RKmax are not "arbitrary parameters" and cannot be set to any value you like. Too large "effective" RKmax

Re: [Wien] LAPW2: semicore band-ranges too large

2015-02-15 Thread delamora
Thank you, after I sent the message I tried > Ru NPT= 781 R0=0.1 RMT= 1.92Z: 44.0 > O NPT= 781 R0=0.00010 RMT= 1.84 > ATOM -3: X=0.5000 Y=0. Z=0. >-3: X=0. Y=0.5000 Z=0. Z: 8.0 And things went much better