Luc Berger
Luc Berger
@jczhang07 @seanofthemillers We are okay with this PR in general but we would like to save the temporary data in the KokkosKernels handle or the KokkosKernels controls instead of having...
@seanofthemillers the shared pointer can be stored in the handle or controls of Kokkos Kernels then, although I have to say having Tpetra destroy and create the matrix on the...
I believe @brian-kelley is already looking into this, @e10harvey maybe you can coordinate to work around that for Gauss-Seidel?
@ndellingwood should we close this?
FYI, this should have been fixed now, in develop at least and will land in 4.0.0. I am not sure if it made it for 3.7.1?
Okay, so I think we can close this now, thanks @brian-kelley and @ndellingwood for tracking this : )
I think so but @jgfouca should probably confirm first.
Reading old issues before the end of the year so I see we did not fully finish this conversation. Here is the approach I recommend: 1. do a coarse grain...
Closing this as it was superseded by #1754, #1896 and #1908
At this point we are testing HIP in our CI, everything is building correct : )