Refactor MGR coarse level matrix computation - Githubissues

hypre-space / hypre

Parallel solvers for sparse linear systems featuring multigrid methods.

https://www.llnl.gov/casc/hypre/

Other

674 stars 185 forks source link

Refactor MGR coarse level matrix computation #1102

Closed victorapm closed 2 months ago

victorapm commented 3 months ago

Main changes are:

Simplify hypre_MGRSetup implementation
- Add HYPRE_MGRSetNonGalerkinMaxElmts to control truncation in MGR's non-galerkin option
Fix MGR's non-galerkin truncation for max_elmts == 0 (no truncation)
Add dedicated file par_mgr_rap.c for coarse level matrix computation routines
Add restriction operator info to MGR statistics

Additional changes that follow up:

Add hypre_CSRMatrixTruncateDiag to truncate a matrix to its diagonal
Add utilities/stl_ops.c for STL-like operations valid for host and device.