| d7059042 | 27-Nov-2016 |
stefano_zampini <stefano.zampini@gmail.com> |
KSPFETIDP: added default identity for pressure preconditioner
With uniform grids, the default pressure preconditioner for Stokes can be obtained as
-fetidp_pmass_ksp_type richardson -fetidp_pmass_k
KSPFETIDP: added default identity for pressure preconditioner
With uniform grids, the default pressure preconditioner for Stokes can be obtained as
-fetidp_pmass_ksp_type richardson -fetidp_pmass_ksp_richardson_scale xx -fetidp_pmass_pc_type none -fetidp_pmass_ksp_max_it 1
with xx = h^-3 (see eq. 6.6 Xuemin's paper)
show more ...
|
| af140850 | 27-Nov-2016 |
stefano_zampini <stefano.zampini@gmail.com> |
PCBDDC/KSPFETIDP: don't restrict to alldofs\setminus interfacepressure
just zero rows and columns of Amat relative to interface pressure
fix a couple of bugs with Pre/PostSolve flags
It works with
PCBDDC/KSPFETIDP: don't restrict to alldofs\setminus interfacepressure
just zero rows and columns of Amat relative to interface pressure
fix a couple of bugs with Pre/PostSolve flags
It works with incompressible Stokes with Taylor-Hood; there are still problems when using the change of basis
show more ...
|
| f5f0d762 | 26-Nov-2016 |
Barry Smith <bsmith@mcs.anl.gov> |
PCReset_FieldSplit() did not work correctly. Do a simple reset that cleans out the old data structures
The drawback to this is that the options are lost to the inner KSP objects; but that is better
PCReset_FieldSplit() did not work correctly. Do a simple reset that cleans out the old data structures
The drawback to this is that the options are lost to the inner KSP objects; but that is better than
buggy code. Eventually we should try to preserve the inner KSP, while removing all the matrices, vectors, and IS
Patch was provided by reporter.
Time: .5 hours
Reported-by: Sander Arens <Sander.Arens@ugent.be>
show more ...
|