Jed is right about the numerics. However, this does not look hard. Here is
my try at it:
https://gitlab.com/petsc/petsc/-/merge_requests/4838
Please tell me if this works and I will make a test and merge.
Thanks,
Matt
On Thu, Feb 10, 2022 at 6:47 PM Jed Brown wrote:
> Susanne, do
Hi Mark,
Thanks for the information.
@Junchao: Given that there are known issues with GPU aware MPI, it might be
best to wait until there is an updated version of cray-mpich (which hopefully
contains the relevant fixes).
Thank You,
Sajid Ali (he/him) | Research Associate
Scientific Computing
Sounds good. Note that if you use direct solvers, that extra node is basically
free because the vertex separators are unchanged. It's a marginal cost in the
storage of assembled matrices and the length of state vectors. And in 3D, even
less significant.
Susanne Claus writes:
> Dear Matthew
Dear Matthew and Jed,
Brilliant. Thank you so much!
Your changes work like a charm Matthew (I tested your branch on the gmsh
file I sent) and thank you so much for your advice Jed. The loss of one
order of convergence for an inf-sup stable pressure discretization seems
indeed a very high