[Trilinos-Users] 2D partitioning on sparse matrix vector multiplication

Da Zheng zhengda1936 at gmail.com
Thu Jan 21 09:29:29 EST 2016


Sparse matrix vector multiplication in Tpetra or Epetra by default
uses 1D partitioning on the sparse matrix for parallelization. What is
the right way to do 2D partitioning on the sparse matrix? I guess we
need to use Zoltan to do things like that. Hopefully, it's not very
complex. Is there any example code for this?


More information about the Trilinos-Users mailing list