[lammps-users] fix thermal/conductivity

Hi,
I’m using the fix thermal/conductivity feature of LAMMPS to calculate the thermal conductivity of a molecular crystal with triclinic symmetry.
My question is when the simulation box is divided into Nbin layers in the edim (x, y, or z) direction, the parallelepiped region of the simulation box is divided by planes perpendicular to the edim direction or parallel to the box faces?

Another question is according to the manual, only the ewald/n style can be used with non-orthoganal simulation boxes, but ewald/n is much slower than pppm, is it possible to implement pppm with non-orthoganal boxes? How hard is it?

Thanks.

Dong

  I'm using the fix thermal/conductivity feature of LAMMPS to calculate the
thermal conductivity of a molecular crystal with triclinic symmetry.
My question is when the simulation box is divided into Nbin layers in the
edim (x, y, or z) direction, the parallelepiped region of the simulation box
is divided by planes perpendicular to the edim direction or parallel to the
box faces?

Parallel to box faces.

Another question is according to the manual, only the ewald/n style can be
used with non-orthoganal simulation boxes, but ewald/n is much slower than
pppm, is it possible to implement pppm with non-orthoganal boxes? How hard
is it?

Hard. I don't know how to do it. Can't recall if I've seen a paper
discussing how to do it either.

Steve