[Wien] runsp -it: lapw1 called thrice?

[Peter Blaha]

Yes, this can happen.

Towards end of the scf (or when mixer mixes very little, the iterative 
diagonalization may fail with a linear dependency error (Cholesky). This 
is trapped in the run_lapw script and a "normal" lapw1 is restarted.

If it happens more often, one should change the number of bands in the 
iterative diagonalization (NBAND in the bottom of case.in1).

The "WARNINGS" probably come from mixer, where he finds that some mixing 
conditions are strange and issues a warning. Unfortunately, no explicit 
explanation of this warning is printed.



On 10/30/2015 12:03 PM, Elias Assmann wrote:
> Hi List,
>
> I started a calculation with ‘runsp … -it -noHinv -min’, and noticed
> that lapw1 was being called three times per cycle: once with ‘-up
> -it’, then ‘-up’, ‘-dn’ without ‘it’ (after the first two, when :MIX
> switched from PRATT to MSE1a).  Is that normal?
>
> Note that the calculation seemed to be running normally; there was a
> *WARNING** from scfm, but I could not find any details about that.
>
>
> 	Elias
>
>
>  From the scf:
>
> :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -222788.07263654
> :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -222788.08807566
> :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -222788.07629953
> :ENE  : *WARNING** TOTAL ENERGY IN Ry =      -222788.10757386
>
> :DIS  :  ( 0.0516339 for atom    5 spin 1)      0.0115736
> :DIS  :  ( 0.2895640 for atom    1 spin 1)      0.0590266
> :DIS  :  ( 0.1109556 for atom    1 spin 1)      0.0266441
> :DIS  :  ( 0.4511020 for atom    2 spin 2)      0.0847330
>
> Here is the :log file (first column is the runtime, I removed the date
> for brevity):
>
>           >   (runsp) options: -p -ec .0001 -cc .0001 -fc 0.5 -i 1000 -it
> -noHinv -orb -min
> 00:00:28 > (x) lapw0 -p
> 00:00:00 > (x) orb -up -p
> 00:00:00 > (x) orb -dn -p
> 00:02:01 > (x) lapw1 -it -up -p -orb -noHinv -c
> 00:01:59 > (x) lapw1 -it -dn -p -orb -noHinv -c
> 00:00:44 > (x) lapw2 -up -p -c -orb
> 00:00:45 > (x) sumpara -up -d
> 00:00:38 > (x) lapw2 -dn -p -c -orb
> 00:00:44 > (x) sumpara -dn -d
> 00:00:41 > (x) lapwdm -up -p -c
> 00:00:00 > (x) sumpara -up -d
> 00:00:42 > (x) lapwdm -dn -p -c
> 00:00:01 > (x) sumpara -dn -d
> 00:00:01 > (x) lcore -up
> 00:00:02 > (x) lcore -dn
> 00:00:11 > (x) mixer -orb
> 00:00:28 > (x) lapw0 -p
> 00:00:00 > (x) orb -up -p
> 00:00:25 > (x) orb -dn -p
> 00:01:06 > (x) lapw1 -it -up -p -orb -noHinv -c
> 00:01:07 > (x) lapw1 -it -dn -p -orb -noHinv -c
> 00:00:44 > (x) lapw2 -up -p -c -orb
> 00:00:45 > (x) sumpara -up -d
> 00:00:45 > (x) lapw2 -dn -p -c -orb
> 00:00:45 > (x) sumpara -dn -d
> 00:00:43 > (x) lapwdm -up -p -c
> 00:00:00 > (x) sumpara -up -d
> 00:00:41 > (x) lapwdm -dn -p -c
> 00:00:00 > (x) sumpara -dn -d
> 00:00:01 > (x) lcore -up
> 00:00:03 > (x) lcore -dn
> 00:00:12 > (x) mixer -orb
> 00:00:28 > (x) lapw0 -p
> 00:00:00 > (x) orb -up -p
> 00:00:28 > (x) orb -dn -p
> 00:01:08 > (x) lapw1 -it -up -p -orb -noHinv -c
> 00:01:59 > (x) lapw1 -up -p -orb -c
> 00:02:01 > (x) lapw1 -dn -p -orb -c
> 00:00:39 > (x) lapw2 -up -p -c -orb
> 00:00:44 > (x) sumpara -up -d
> 00:00:39 > (x) lapw2 -dn -p -c -orb
> 00:00:44 > (x) sumpara -dn -d
> 00:00:40 > (x) lapwdm -up -p -c
> 00:00:01 > (x) sumpara -up -d
> 00:00:40 > (x) lapwdm -dn -p -c
> 00:00:01 > (x) sumpara -dn -d
> 00:00:01 > (x) lcore -up
> 00:00:02 > (x) lcore -dn
> 00:00:14 > (x) mixer -orb
> 00:00:27 > (x) lapw0 -p
> 00:00:00 > (x) orb -up -p
> 00:00:23 > (x) orb -dn -p
> 00:01:07 > (x) lapw1 -it -up -p -orb -noHinv -c
> 00:02:01 > (x) lapw1 -up -p -orb -c
> 00:02:01 > (x) lapw1 -dn -p -orb -c
> 00:00:35 > (x) lapw2 -up -p -c -orb
> 00:00:45 > (x) sumpara -up -d
> 00:00:33 > (x) lapw2 -dn -p -c -orb
> 00:00:45 > (x) sumpara -dn -d
> 00:00:40 > (x) lapwdm -up -p -c
> 00:00:01 > (x) sumpara -up -d
> 00:00:40 > (x) lapwdm -dn -p -c
> 00:00:01 > (x) sumpara -dn -d
> 00:00:01 > (x) lcore -up
> 00:00:02 > (x) lcore -dn
>           > (x) mixer -orb
>

-- 

                                       P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300             FAX: +43-1-58801-165982
Email: blaha at theochem.tuwien.ac.at    WIEN2k: http://www.wien2k.at
WWW:   http://www.imc.tuwien.ac.at/staff/tc_group_e.php
--------------------------------------------------------------------------