AFD-Illinois / iharm3d

Hybrid MPI/OpenMP 3D HARM with vectorization
Other
12 stars 9 forks source link

Conducting atmosphere test #43

Closed vedantdhruv96 closed 2 years ago

vedantdhruv96 commented 2 years ago

Converges at second order (up to 256^2). Errors in q flatten out at higher resolutions at ~10^(-10). Finite differencing artifact since we require the connection coefficients to compute the Eckart heat flux?

Is this the best we can do? Convergence plot in grim considers 512^2 but grim shows larger base errors, due to which the L1 norm at 512^2 ~10^(-8)