Mesh size largly differes in 'parabolic' and 'pdenonlin' functions to give a solution.

Hi,

I'm going to solve an elliptic PDE where the gradient of solution is extremely large in a specific portion of the domain. 'pdenonlin' demands a really fine mesh size around this region and consequently large amount of memory. but when I choose to solve its parabolic PDE version and increase the time large enough to get a steady-state solution, it works fine with larger mesh sizes. So I want to know is this an intrinsic property of these two functions or I'm missing some thing here?

thanks,

[Mohammad Monfared]

This is not necessarily true.