IALGO for MD calculation

[okuno]

Dear Vasp Users and administrator.

I'm now trying to run MD calculation for amorphous system.

I use IALGO=48 which is recommended for large molecular system.
But the calculation does not conversed. However if we switch IALGO= 38, the calculation conversed.

In MD calculation, can we use IALGO=38,?
And How we should converse if I use IALGO=48 ?

This is my INCAR file

PREC=low
ICHARG=2
EDIFF=1e-4
ISPIN=1

ISYM=0
ISIF=2
IBRION=0
POTIM=2.0
NSW=5
!NSW=750
NBLOCK=1
KBLOCK=10

IALGO=48 ! RMM-DIIS method for electrons
NELMIN=4
NELM=100

BMIX = 2.0
MAXMIX=50
TEBEG=300
SMASS=0

ISMEAR=1
SIGMA=0.1
LREAL=Auto

[admin]

the convergence behaviour of IALGO=38 (the Davidson algorithm) might be better for systems with a very dense eigenvalue spectrum. It is recomended to use ALGO=fast, which is a good compromise between calculation speed and safe convergence