[petsc-dev] viewing factored matrices

Jed Brown jedbrown at mcs.anl.gov
Sat Nov 23 14:06:36 CST 2013


Barry Smith <bsmith at mcs.anl.gov> writes:
>    XXSetFromOptions() should never be called where it actually cannot
>    do anything (besides checking lots of strings) because it has
>    already been called on the same configuration. And since the
>    options database hasn’t changed the options will be set but they
>    will be set with the exact same values as before.

Would it be enough to make the options database carry a version number
that was incremented on changes, cached by each object so that they can
quickly short-circuit?  It feels a bit dirty, but would be an easy way
to cut out that processing.  Or it could just be used to audit whether
options were being checked redundantly.
-------------- 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-dev/attachments/20131123/6eaeb01d/attachment.sig>


More information about the petsc-dev mailing list