**The imprimation of TRAN_Poisson for the boundary condition** |
- Date: 2010/08/25 17:24
- Name:
**Yukihiro Okuno**
- Dear OpenMX Developers and Users.
I'm now interesting to the method of NEGF in OpenMX, and see the source code of the determination of Hatree potential in TRAN_Poisson.c
I found in source code of the treatment of boundary condition for Poisson solver by FFT, (TRAN_Poisson_FFT, or TRAN_Poisson_FFT_Extended).
v0 = dVHart_Grid_e[0][j*Ngrid3]+k v1 = dVHart_Grid_e[1][j*Ngrid3]+k
a = (v1-v0)/(x1-x0) b = v0 - a*x0
/*
v = ax + b
*/
v = a*x
dVHart_Grid[MN] += v
Why v=ax+b is commented out, and only add v=ax ?
I think it should be v=ax + b is added in order to adjust the boundary condition
v0 = dVHart_Gride[0] and v1 = dVHart_Grid[1].
Sincerely,
Yukihiro Okuno.
| |