The carbon environment dependent interaction potential (EDIP), often called the EDIP potential (yes, like ATM machine...).
These are the publications that describe the carbo-EDIP potential. If you are using this potential, please consider citing these works.
-
N.A. Marks, "Generalizing the environment-dependent interaction potential for carbon", 2000, Phys. Rev. B 63, 035401, DOI 10.1103/PhysRevB.63.035401
-
N.A. Marks, "Modelling diamond-like carbon with the environment-dependent interaction potential", 2002, J. Phys.: Condens. Matter 14 2901, DOI: 10.1088/0953-8984/14/11/308
Size of arrays:
parameter (NMAX=200000,NNN=50,NP=0.5*NNN*NMAX)
common /ZDERV/ dzdx(NNN,3), dzdxx(NNN,NNN,3)
common /ZFORC/ finiteforce(NNN), finiteforce2(NNN,NNN)
common /DRXYZ/ dr(NMAX,NNN),dx(NMAX,NNN,3),dxdr(NMAX,NNN,3)
common /INEAR/ near(NMAX,NNN),num(NMAX),kron(NNN,NNN)
common /ZCYCL/ f(NMAX,NNN), df(NMAX,NNN), dzz(NMAX,NNN)
common /PAIRS/ ipair(NP),jpair(NP),inum(NP),jnum(NP),npair
NMAX*NNN:
/DRXYZ/ 5*NMAX*NNN (doubles)
/INEAR/ 1*NMAX*NNN (integer)
/ZCYCL/ 3*NMAX*NNN (doubles)
/PAIRS/ 2*NMAX*NNN (doubles)
assuming 8 bytes per value, we have 88*NMAX*NNN bytes
For NMAX=200000 and NNN=50, NMAX*NNN=10 million
For NMAX=200000 and NNN=50, 88*NMAX*NNN=880 million
For NMAX=400000 and NNN=50, 88*NMAX*NNN=1.76 billion (~1.8 GB)
There are several other large arrays, compilation fails with
NMAX=400000, but ok with NMAX=350000. Looks like we have a
maximum memory allocation of 2GB. Pretty small ...