Parallel Krylov adjoint sensitivity analysis example
3D Advection diffusion PDE with homogeneous Neumann B.C.
Computes gradient of G = int_t_Omega ( c_i^2 ) dt dOmega
with respect to the source values at each grid point.

Domain:
   0.000000 < x < 20.000000   mx = 20  npe_x = 2 
   0.000000 < y < 20.000000   my = 40  npe_y = 2 
   0.000000 < z < 20.000000   mz = 20  npe_z = 1 

Begin forward integration... done.   G = 8.232843e+03

Final Statistics.. 

lenrw   = 180946     leniw =    212
llrw    = 180856     lliw  =     80
nst     =    118
nfe     =    125     nfel  =    141
nni     =    121     nli   =    141
nsetups =     17     netf  =      0
npe     =      2     nps   =    234
ncfn    =      0     ncfl  =      0

Begin backward integration... done.

Final Statistics.. 

lenrw   = 361716     leniw =    212
llrw    = 180856     lliw  =     80
nst     =     70
nfe     =     80     nfel  =    133
nni     =     76     nli   =    133
nsetups =     15     netf  =      0
npe     =      2     nps   =    204
ncfn    =      0     ncfl  =      0