Luca Bertagna
Luca Bertagna
It might be that some ranks pick up a high-speed region, while others pick up slow-speed ones. This is especially possible for highly non-uniform meshes.
Well, user guides are key to keep a software library alive after its inception developers retire or move on to other projects. I of course encourage such heroic task, though...
I'm ok with pretty much any strategy that would prevent random failures due to timeout, when the reason for the time out is not on the albany side. Disabling a...
I actually thought that large tests meant also long. Large exodus files should be a good proxy for 'long test', either because of large mesh or large number of timesteps....
I'm all for keeping the (storage) size of tests small. Personally, I think we should not even have 10k+ DOFs tests, and let each project (LandIce, LCM, Aeras, whatnot) have...
What about the rest? I found a problem in ATO (`ATO/problems/LinearElasticityModalProblem.hpp`) that uses the GatherEigenData evaluator, so I guess we may have to keep it. Although, I don't know if...
I stumbled on this issue. Now that pruning of non-LandIce related stuff is happening, this might be relevant. In particular, I would like to ask you guys what's the plan...
I haven't counted them. I'm guessing about 10 or so, mostly in the basic and demo tests. But it's ok, we can wait.
> @mcarlson801, can you please attach the Trilinos configure script, the cmake STDOUT&STDERR, and the CMakeCache.txt file and I will take a look? Otherwise, where is `mpi_usempif08` coming from? @bartlettroscoe...
Ok, thanks!