How to define the starting magnetic moments for each atom? about abacus-develop

[DYang90]

There are 3 input structures could be considered.
1.
symbol='Ni',magmom=2,...
symbol='Ni',magmom=-2,...
symbol='O',magmom=0,...
symbol='O',magmom=0,...
2.
symbol='Ni Ni O O',
magmom='2 -2 2*0',
....
3.
symbol='Ni1',magmom=2,...
symbol='Ni2',magmom=-2,...
symbol='O',magmom=0,...

The second one might be suitable for manual inputs，because some values of magmom can be abbreviated as ntype*magmom.
The last one is more complex because a number is always needed for classifying the atoms with magnetic moments and Hubbard parameters, eg.
symbol='Co Co Co Co...',
magmom='23 22...',
ldauu='4.5 3.0 4.5 3.0...'
...
=>
symbol='Co1',magmom=3,ldauu=4.5...
symbol='Co2',magmom=3,ldauu=3.0...
symbol='Co3',magmom=2,ldauu=4.5...
symbol='Co4',magmom=2,ldauu=3.0...
...

[mohanchen]

OK, we will try to update the function.

[DYang90]

OK, we will try to update the function.
I suggest that the magnetic moments should be specified in the STRU file. In this way, we can edit the magnetic moments in STRU file by some visualizer tools.

[ZhouGD2000]

I use the second structure suggested by @DYang90, which is also used in VASP, but also preserve the original way.
The input struture is

magmom m1  m2 n3*m3 

where m1,m2,m3,....following the same order as the atom type in STRU file, and is defined in INPUT file. There must be NO space before and after *.

Either defining the magnetic moment in this way or in STRU file is allowed, but not both of them.
NOTICE: this is just a modification in the input structure.

[DYang90]

I use the second structure suggested by @DYang90, which is also used in VASP, but also preserve the original way.
The input struture is

magmom m1  m2 n3*m3 

where m1,m2,m3,....following the same order as the atom type in STRU file, and is defined in INPUT file. There must be NO space before and after *.

Either defining the magnetic moment in this way or in STRU file is allowed, but not both of them.
NOTICE: this is just a modification in the input structure.

Perhaps some forethought for noncollinear magnetic case is necessary.

[wenfei-li]

It seems this issue has been resolved, therefore it will be closed