Will Pazner
Will Pazner
Hi @Sbozzolo, I don't think anything significant has changed in the ParaView output since then. Perhaps use of Umpire is changing the performance. It would be useful if you could...
Yes, my current suspicion is that this is Umpire-related. My guess is that if you measure with each of the two commits, with and without Umpire, you will see that...
Some related discussion here: https://github.com/mfem/mfem/issues/3965#issuecomment-1802931258 What do you think?
Just leaving some references to past discussions on smart pointers: * https://github.com/orgs/mfem/discussions/4614 * #4629 Yeah, it's pretty fundamental that `shared_ptr` can't release ownership. It probably makes reasoning about them easier....
I think this is handled in a more general way in #4028.
Thanks @ebchin. Do you have a reproducer?
This PR is now [under review](https://github.com/mfem/mfem/blob/master/CONTRIBUTING.md#pull-requests) (see the table in the PR description). To help with the review process, please **do not force push** to the branch.
> @tzanio The sole failure at the CI level (Tests / macos-latest-make-dbg-par-int32-fp64, and more precisely at the "Xcode version setup (MacOS)" step) does not seem to have anything to do...
Hi @Wayne901, If you call `CGSolver::SetOperator` and there is already a preconditioner added to the CG solver, it will in turn call `SetOperator` on the preconditioner. `LORSolver::SetOperator` ends up calling...