Phil Miller - NOAA comments

Results 319 comments of


                                            Phil Miller - NOAA

Sacado: Invalid FAD dimension

That change was motivated by both consistency with the behavior of View, and benefits to our ability to implement routines like `create_mirror` and `deep_copy` correctly, including upcoming improvements to striding...

Sacado: Invalid FAD dimension

If you post your desired modification to `as_view_of_rank_n`, I can try to review it quickly

Sacado: Invalid FAD dimension

Yeah, I'd approve that. While we're at it, are there any other functions you can see that need this treatment added?

Revise Public and Private Header Files in Kokkos

3.7.00 is shipped, with the relevant guarding macros. I think this can be closed, but the documentation needs to be updated. I just opened an issue for that

Trilinos compilation errors, intel/19 + OMP: internal compiler error 101003_1112 with Sacado, Intrepid2 unit tests

Workaround was shipped in 3.7.00

Clang/10+Cuda test failure: cuda.reducers_complex_double cudaStreamSynchronize(stream) error( cudaErrorIllegalAddress): an illegal memory access was encountered

Fix released in 3.7.00

[WIP] Add m_stride to layout to support propagation of Padding

Earlier versions of this PR explicitly modified `Kokkos_DynRankView.hpp` as well - is that no longer necessary? Could we get some testing to make sure that it's covered too?

[WIP] Add m_stride to layout to support propagation of Padding

Both failures in Jenkins now are actually timing-based, and unrelated to this PR

[WIP] Add m_stride to layout to support propagation of Padding

The merge looks good to me

Kokkos::sort breaks on HIP with certain integer arrays

Fixed in the 3.7.00 release