Eric Phipps

@ndellingwood This PR includes a change in Kokkos. How should that be handled?

> @ndellingwood @etphipp The fix in Kokkos was already submitted to Kokkos more than a month ago in > > * [complex(math) : log10 support kokkos/kokkos#5009](https://github.com/kokkos/kokkos/pull/5009)... So this change is...

I guess my question is, since the changes are already in kokkos-develop, can we not just merge this PR including its kokkos changes, which will then be overwritten at the...

So I guess we just wait for the next Kokkos snapshot then.

Unfortunately, I don't think there has been any attempt to make Stokhos work with UVM disabled. It is unclear how much work would be involved in making it work without...

I think at the minimum we would just have to disable all of the PCE stuff. It works similarly to DFad in Sacado, which is pretty much impossible to make...

When I get a chance I will try and figure out a more fine-grained set of tests that need to be disabled or refactored.

@trilinos/framework I am working on resolving these failures related to UVM. However, some of these tests appear to already get disabled on at least some architectures and not others and...

> @etphipp If you're using the GenConfig tool, the tests that get disabled are in `packages/framework/ini-files/config-specs.ini`. I'm not using GenConfig. I added a print statement right before the call to...

@trilinos/tpetra I am trying to get these tests to work without UVM, but am running into issues with `Tpetra::CrsGraph::getLocalRowView()` throwing exceptions that a device view is still alive. There are...