Replace velocity for slowness squard as parameter in the wave equation

[joao-bapdm]

• Use the squared slowness $m$ formulation

  $m\frac{\partial^2 p}{\partial t^2} + \eta \frac{\partial p}{\partial t} - \nabla^2 p = f$.

  instead of the velocity $c$.

  $\frac{\partial^2 p}{\partial t^2} + \eta \frac{\partial p}{\partial t} - c^2 \nabla^2 p = f$.

• A missing factor of 2 was added to the denominator of the time derivative in the damping term of the absorbing boundary layer.

• The test_solution.py apparently uses an old reference produced with simwave itself, and maybe could be revised.

[krober10nd]

What's the benefit of this new formulation?

[joao-bapdm]

Having $p \cdot \partial_{tt}{p}^\dagger$ in the gradient instead of $\nabla p \cdot \nabla p^\dagger$

[jaimesouza]

@joao-bapdm Could you update the 3D as well in https://github.com/HPCSys-Lab/simwave/blob/master/simwave/kernel/backend/c_code/forward/constant_density/3d/wave.c ?