no convergence

Posted on : November 02, 2016 (Wed) 11:50:01

by Wang Busheng

Dear all, I always get the following result with no convergence when I do the scf calculation of Pt.

My input:
c----------------------CrAs----------------------------------
go data/Pt
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
fcc 2.0744 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 1.0 nrl mjw mag init
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
update 4 500 0.03
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
Pt 1 1 0.0 3 78 100
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0 0 0 Pt
c------------------------------------------------------------

And my output:
-------------------------------------------------------------------------------
.
.
.
itr= 98 neu= -28.3469 moment= 0.0099 te= -37135.45180991 err= -11.306
itr= 99 neu= -28.0681 moment= -0.7258 te= -37132.00132853 err= -11.378
itr=100 neu= -28.0708 moment= -0.6782 te= -37132.63099059 err= -11.921
*** no convergence
interval=100 cpu time= 63.91 sec
sdftyp=mjw reltyp=sra dmpc= 0.03000
Pt
itr=100 neu-28.0708 chr,spn 4.8344 -0.6782 intc,ints 28.5752 -0.66

rms err= -2.125 -1.921
ef= 3.0128443 3.0621557 def= 2.0008462 16.1030052
total energy= -37132.630990594
------------------------------------------------------------------------------------
Is there anything wrong with my input file ?

Thank you for reply and give me your views.

[Re:06] no convergence

Posted on : November 04, 2016 (Fri) 21:32:05

by Wang Busheng

Dear Administrator,

It woks well, thank you so much!

[Re:05] no convergence

Posted on : November 04, 2016 (Fri) 17:42:41

by Administrator

Use the following input:

c----------------------CrAs----------------------------------
go data/Pt95Fe5
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
fcc 7.3793 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 1.0 nrl mjw mag init
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
update 4 500 0.02
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
PtFe 2 1 0.0 3 78 95
26 5
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0 0 0 PtFe
c------------------------------------------------------------

The point is using a little bit smaller value for "lmix". Also the value of "ncmp" should be 2.

KKR Admistrator

[Re:04] no convergence

Posted on : November 04, 2016 (Fri) 17:29:14

by Wang Busheng

Dear KKR Administrator,

I have fixed the lattice constant in my input file, but there is still a '*** no convergence' in Fe0.05Pt0.95 scf-calculation. the input as blow:

c----------------------CrAs----------------------------------
go data/Pt95Fe5
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
fcc 7.3793 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 1.0 nrl mjw mag init
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
update 4 500 0.03
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
PtFe 1 1 0.0 3 78 95
26 5
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0 0 0 PtFe
c------------------------------------------------------------

But I get no '*** no convergence’ issue with either Fe0.1Pt0.9 or Fe0.15Pt0.85.

Any comments or documentation for this problem will be greatly appreciated.

[Re:03] no convergence

Posted on : November 02, 2016 (Wed) 16:44:30

by Wang Busheng

Dear KKR Administrator,

Oh, thank you so much for your reminding!

My mistake with the unit convert.

[Re:02] no convergence

Posted on : November 02, 2016 (Wed) 14:01:01

by Administrator

Dear Wang Busheng,

the lattice constant of fcc Pt is around 7.41 (a.u.). Note that the lattice constant should be given in the unit of Bohr (0.529177 A).

KKR Administrator

[Re:01] no convergence

Posted on : November 02, 2016 (Wed) 12:57:02

by Wang Busheng

sorry, the output is corresponding to input in which 'maxitr' is 100, not 500.