Lagrange multiplier for Dirichlet boundary condition

General Discussion
16

Dear @prj @marchywka , sorry to bother you again…
I have now used ChangeNumbering() to converge the FreeFem dimension to PETSc dimension as follows. However, the error is still there


 f(0:undof + pndof - 1) = 0;
//f(undof + pndof : tndof - 1) = Af;
 real[int] Afnew;
ChangeNumbering(AA, Afx [], Afnew);
 f(undof + pndof : tndof - 1) = Afnew;

us(undof:tndof - 1) = 2.; 
real[int] ufnew;
 ChangeNumbering(UU, ufx [], ufnew);
us(0 : undof - 1) = ufnew;

When I use one core to run my code, there is no error and the result is correct, but there is a problem of using two cores.
stokes_lagrange_parallel_1.edp (3.4 KB)

The error looks like:

image
image1007×632 36.1 KB

BTW, what was the command to output the error message as you showed above? I use
ff-mpirun -n 2 stokes_lagrange_parallel_1.edp -v 10

Thanks very much for your help.