Atomic magnetic information loss in STRU file during the dpgen + abacus workflow #3942
Closed
16 tasks
Labels
Bugs
Bugs that only solvable with sufficient knowledge of DFT
Describe the bug
When I perform a workflow of dpgen + abacus, I prepare a perturbed structure in init/STRU.01x01x01/00.place_ele/sys-0036-0096/STRU file and add atomic magnetic information in STRU file. I use this structure as a initial structure to use dpgen run param.json machine.json. However, when I reach the 02.fp step, it seems that after a dpdata conversion, the atomic magnetic information in the STRU file has been lost.
Expected behavior
I am not sure this problem can be attributed to ABACUS or Dpgen. But I wish that atomic magnetic information don't loss.
To Reproduce
my initial STRU file:
ATOMIC_POSITIONS
Direct //Cartesian or Direct coordinate.
Nb // Element type
0.0 // magnetism
36 // number of atoms
0.803569973 0.267859995 0.082450002 mag 0.5
0.303570002 0.767859995 0.082450002 mag 0.5
0.196429998 0.398809999 0.250880003 mag -0.5
0.696430027 0.898810029 0.250880003 mag -0.5
............
STRU in 02.fp file:
ATOMIC_POSITIONS
Cartesian # Cartesian(Unit is LATTICE_CONSTANT)
Nb
0.0
36
6.152880000000 3.620300000000 3.388680000000 1 1 1
2.277500000000 10.205400000000 3.441630000000 1 1 1
1.473130000000 5.270050000000 10.539500000000 1 1 1
5.311980000000 11.826200000000 10.575700000000 1 1 1
1.461040000000 3.549240000000 3.348810000000 1 1 1
..........
It is obvious that atomic magnetic information have been lost.
This phenomenon can be observed from the file in abacus-user-guide/examples/abacus-dpgen/
Environment
No response
Additional Context
No response
Task list for Issue attackers (only for developers)
The text was updated successfully, but these errors were encountered: