Worse, validity of the cache changes whether it has a barrier (in debug
mode) or not. I would prefer to avoid calling PetscCommDuplicate() because
it is fewer key lookups and then we can make PetscCommGetThreadComm
explicitly NOT collective. Any objections?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: 
<http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20121020/95eafecb/attachment.html>

Reply via email to