Jie,
It turns out the SparseILU needs to be tuned very carefully. I set the
AdditioalData to (0.03, 0) the solver now works fine. Although it is still
slower than UMFPACK. Maybe using just the right parameters in AdditionalData
can further improve it. I observed similar thing in step-57 too. Using ILU +
GMRES for A inverse won't even converge in reasonable time with default
AdditionalData. Playing with the AdditionalData helps... In a word, SparseILU
seems to be sensitive to the AdditionalData and how to choose the right
parameter is unclear.
Yes, that's what I meant to say in my last email: the ILU is quite sensitive
to choosing parameters.
Do you have a set of parameters that works significantly better than what we
currently use in step-57? Would you like to propose we make changes to step-57
in this direction?
Wolfgang, would you like me to contribute this code to the code gallery? Is
there anything in this code that I should polish?
You mean your time-dependent N-S solver? Yes, we would absolutely love to have
it as part of the code gallery! If you contribute it as a pull request as
explained on the website, we can take a look and see what should/needs to be
polished, though we're typically quite relaxed about code gallery programs!
Best
Wolfgang
--
------------------------------------------------------------------------
Wolfgang Bangerth email: [email protected]
www: http://www.math.colostate.edu/~bangerth/
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
