On 2/14/23 19:04, 'yy.wayne' via deal.II User Group wrote:
Yes the diagonal could be negative when k rises. Will Chebyshev preconditioner
smoother converge
when there are negative entries?
I have no idea :-)
In my case k hasn't rise so large that diagonal entries is negative. The
negative entries shows up only
when compute_normal_flux constraints is applied. I think
AffineConstraints::distribute_local_to_global
<https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.dealii.org%2Fcurrent%2Fdoxygen%2Fdeal.II%2FclassAffineConstraints.html%23a373fbdacd8c486e675b8d2bff8943192%3A~%3Atext%3DThus%2C%2520if%2520a%2520degree%2Cin%2520the%2520global%2520matrix.&data=05%7C01%7CWolfgang.Bangerth%40colostate.edu%7Cd2ecd8f014fc4c51b63308db0ef8fd28%7Cafb58802ff7a4bb1ab21367ff2ecfc8b%7C0%7C0%7C638120234786119642%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=c2TyCwiADwVWxKmtOEDrujHEkVC%2F8QPR5k7gIip3OoQ%3D&reserved=0>
is the cause since in matrix-based assemble, this function add a positive
value on the constrained
dofs while matrix-free does not. Therefore the diagonals are different between
matrix-free and matrix-based.
That's possible. I don't know enough about the matrix-free framework to tell
one way or the other, but maybe some of my colleagues here will know.
Best
W.
--
------------------------------------------------------------------------
Wolfgang Bangerth email: bange...@colostate.edu
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 dealii+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/dealii/92c9d94b-6637-8d4a-e883-f47e2fc8d0ec%40colostate.edu.
