1. 4
  1.  

  2. 1

    I always thought this might be the case; for any non-trivial CFD simulation, it always seemed that the solution convergence was fragile as hell. Nice to see some data to back up my intuition.