[gmx-users] Hessian matrix within gmx mdrun

[Adriaan Riet]

Hello,

I'm trying to do a normal modes analysis on a large crystalline system
(~100,000 atoms of MgO). I expected the hessian to be (300,000x300,000), or
NDIM*natoms, but by poking around in the source code a *little*, I can see
that the Hessian is made only by the atom types, not by the actual atoms.
Is there a way to do it based on individual atoms? I've tried looking for
examples on how to do this, but all the links I can find to details are
broken.

Thanks,
Adriaan Riet