position_step

torch_sim.integrators.md.position_step(state, dt)

Update particle positions using current velocities.

This function performs the position update step of velocity Verlet integration by propagating particles according to their velocities over timestep dt. It implements the equation: r(t+dt) = r(t) + v(t) * dt

Parameters:

• state (MDState) – Current system state containing positions and velocities

• dt (torch.Tensor) – Integration timestep, either scalar or with shape [n_systems]

Returns:

Updated state with new positions after propagation

Return type:

MDState