[Wien] SIGSEGV, segmentation fault in lapw1: case of 32 atoms, while no errors on lighter cases

[Peter Blaha]

Do not play with GMAX, keep it at 14 (at least).

However, you may reduce RKMAX (case.in1, default=7.0), in particular when 
you have small H spheres in your system. Try RKmax=3-4 and set the sphere 
radii according to the tips on the faq page of www.wien2k.at

The segmentation fault is mostlikely due to some computer limitations or 
some limits, which should be set to unlimited.

 > I got a segmentation fault during lapw1. My diagnostics follows. > 
> I run WIEN2k on a case of 32 atoms (an organic molecule), for this time I set Gmax to 14, 
> Gmin is calculated to 8,1632etc
> 
> after about 5-15 s of lapw1 
> the program stops and 
> LAPW1.error says no more than: "error in lapw1"
> and 
> STDOUT:
> forrtl: severe (174): SIGSEGV, segmentation fault occurred
> Image              PC        Routine            Line        Source             
> lapw1c             08076384  Unknown               Unknown  Unknown
> lapw1c             0808BD92  Unknown               Unknown  Unknown
> lapw1c             00000032  Unknown               Unknown  Unknow
> 
> Stack trace terminated abnormally.
> 
> This happens while calculation of 2 following lighter cases works nice:
> These lighter cases are:
> 1. The same structure, 32 atoms. I set Gmax to 10 and choose core orbitals separation 
> energy in lstart to -4 (instead of -6, recomennded) /average lapw1 execution time here is 12 min/
> 2. An analogous structure consisting of 22 atoms -all other conditions are 
> as they are in the error-producing case-/average lapw1 execution time here is 5 min/
> 
> I have IntelP4 2,8 GHz, Goto libs, and 1Gb of Ram 
> of which about 200Mb is free during the calculations
> 
> Please help:
> Can one see which part of a computer such segmentation error concerns?
> Is that so that WIEN stops on an error for heavier cases instead of just working more slow?
> Please send any other of your comments on my story
> 
> Alexei Komolov


                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------