ALGO on magnetic moment?
Posted: Fri Oct 23, 2020 3:41 pm
Hi, everyone,
I ran two separate calculations from the same POSCAR, KPOINTS, POTCAR. The difference in INCAR is small, as one uses ALGO= Normal, and one uses ALGO= Fast. some different value in amix, bmix, amix mag, bmix mag.
Surprisingly, at the first ionic step,, both converged to 1e-4, giving totally different magnetic moment
run 1. (ALGO=Normal)
1 F= -.10838758E+04 E0= -.10838596E+04 d E =-.108388E+04 mag= 0.7993
run 2 (ALGO=Fast)
1 F= -.10842798E+04 E0= -.10842624E+04 d E =-.108428E+04 mag= 0.0518
And further structural relaxation leads to different geometries (energy difference of 0.7 eV !!!!!), as
run 1. (ALGO=Normal)
12 F= -.10838768E+04 E0= -.10838598E+04 d E =-.494185E-03 mag= 0.6096
run 2 (ALGO=Fast)
27 F= -.10843214E+04 E0= -.10843075E+04 d E =-.129200E-03 mag= -1.0044
Any comments is highly appreciated!
with regards
Kai
Details of two INCARs as given below:
INCAR #1
PREC = Normal
ENCUT = 400.000
IBRION = 2
NSW = 100
ISIF = 2
NELMIN = 2
EDIFF = 1.0e-04
EDIFFG = -0.02
VOSKOWN = 1
NBLOCK = 1
NWRITE = 1
NELM = 100
ALGO = Normal (blocked Davidson)
# MAXMIX = 50 #
ISPIN = 2 #1 # no spin polarized calculation
INIWAV = 1
ISTART = 0
ICHARG = 2
LWAVE = .FALSE.
LCHARG = .FALSE.
ADDGRID = .FALSE.
ISMEAR = 0 # recommended for semiconductor
SIGMA = 0.05 # recommended for semiconductor
LREAL = A # Manual recommends A for mixed ions or more than 20 atoms
POTIM = 0.1
AMIX = 0.2 #set amix
BMIX = 0.0001 #almost zero in initial mixer
AMIX MAG = 0.8
BMIX MAG = 0.0001 #almost zero, but 0 will crash some versions
# MAGMOM= 1*1.5 128*0 # SPIN NUMBER*SPIN
INCAR #2
PREC = Normal
ENCUT = 400.000
IBRION = 2
NSW = 200
ISIF = 2
NELMIN = 2
EDIFF = 1.0e-04
EDIFFG = -0.02
VOSKOWN = 1
NBLOCK = 1
NWRITE = 1
NELM = 100
ALGO = Fast #VeryFast # Normal (blocked Davidson)
# MAXMIX = 50 #
ISPIN = 2 #1 # no spin polarized calculation
INIWAV = 1
ISTART = 0
ICHARG = 2
LWAVE = .FALSE.
LCHARG = .FALSE.
ADDGRID = .FALSE.
ISMEAR = 0 # recommended for semiconductor
SIGMA = 0.05 # recommended for semiconductor
LREAL = A # Manual recommends A for mixed ions or more than 20 atoms
POTIM = 0.1
AMIX = 0.1 #set amix
BMIX = 0.0001 #almost zero in initial mixer
I ran two separate calculations from the same POSCAR, KPOINTS, POTCAR. The difference in INCAR is small, as one uses ALGO= Normal, and one uses ALGO= Fast. some different value in amix, bmix, amix mag, bmix mag.
Surprisingly, at the first ionic step,, both converged to 1e-4, giving totally different magnetic moment
run 1. (ALGO=Normal)
1 F= -.10838758E+04 E0= -.10838596E+04 d E =-.108388E+04 mag= 0.7993
run 2 (ALGO=Fast)
1 F= -.10842798E+04 E0= -.10842624E+04 d E =-.108428E+04 mag= 0.0518
And further structural relaxation leads to different geometries (energy difference of 0.7 eV !!!!!), as
run 1. (ALGO=Normal)
12 F= -.10838768E+04 E0= -.10838598E+04 d E =-.494185E-03 mag= 0.6096
run 2 (ALGO=Fast)
27 F= -.10843214E+04 E0= -.10843075E+04 d E =-.129200E-03 mag= -1.0044
Any comments is highly appreciated!
with regards
Kai
Details of two INCARs as given below:
INCAR #1
PREC = Normal
ENCUT = 400.000
IBRION = 2
NSW = 100
ISIF = 2
NELMIN = 2
EDIFF = 1.0e-04
EDIFFG = -0.02
VOSKOWN = 1
NBLOCK = 1
NWRITE = 1
NELM = 100
ALGO = Normal (blocked Davidson)
# MAXMIX = 50 #
ISPIN = 2 #1 # no spin polarized calculation
INIWAV = 1
ISTART = 0
ICHARG = 2
LWAVE = .FALSE.
LCHARG = .FALSE.
ADDGRID = .FALSE.
ISMEAR = 0 # recommended for semiconductor
SIGMA = 0.05 # recommended for semiconductor
LREAL = A # Manual recommends A for mixed ions or more than 20 atoms
POTIM = 0.1
AMIX = 0.2 #set amix
BMIX = 0.0001 #almost zero in initial mixer
AMIX MAG = 0.8
BMIX MAG = 0.0001 #almost zero, but 0 will crash some versions
# MAGMOM= 1*1.5 128*0 # SPIN NUMBER*SPIN
INCAR #2
PREC = Normal
ENCUT = 400.000
IBRION = 2
NSW = 200
ISIF = 2
NELMIN = 2
EDIFF = 1.0e-04
EDIFFG = -0.02
VOSKOWN = 1
NBLOCK = 1
NWRITE = 1
NELM = 100
ALGO = Fast #VeryFast # Normal (blocked Davidson)
# MAXMIX = 50 #
ISPIN = 2 #1 # no spin polarized calculation
INIWAV = 1
ISTART = 0
ICHARG = 2
LWAVE = .FALSE.
LCHARG = .FALSE.
ADDGRID = .FALSE.
ISMEAR = 0 # recommended for semiconductor
SIGMA = 0.05 # recommended for semiconductor
LREAL = A # Manual recommends A for mixed ions or more than 20 atoms
POTIM = 0.1
AMIX = 0.1 #set amix
BMIX = 0.0001 #almost zero in initial mixer