[petsc-users] question about alternative to MatMatSolve

[Jed Brown]

Michael Povolotskyi <mpovolot at purdue.edu> writes:
> In my case each right hand side column is an almost empty vector with a 
> couple of non zero entries.
> Is there any way to use the fact that the right hand side is very sparse 
> and make the calculations faster?

Optimizing for this case would only give a small benefit because the
vector becomes dense during the forward-solve and the entire back-solve
has the dense vector.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20131104/f8d5de91/attachment.pgp>