Tomasetti Romin
Tomasetti Romin
## Enhancement @panzer @rppawlo We think that `Panzer` DOF manager and other classes should use `Teuchos::Comm` instead of the raw MPI type. This would be needed to: * simplify Panzer...
## Summary This PR: * ~~Adds Google Test from a git tag iif tests are enabled~~ Done in #237 * Moves declarations from `kp_space_time_stack.cpp` to `kp_space_time_stack.hpp` to enable users make...
## Summary This PR ensures that the tool `space-time-stack` demangles kernel names. ## Side quests, questions and related issues - [ ] I created new files to store what's common...
As discussed in #218, it would be better if the frame type is an `enum` shared by all tools. See `common/FrameType.hpp` in #218. > I created a new enum FrameType...
As requested by @crtrott in https://github.com/kokkos/kokkos-tools/pull/211#issuecomment-1749332578. The closest to `11.0` in the Docker hub is `11.0.3`.
### Summary As discussed with @dalg24 , the native graph objects should be accessible for advanced users. ### Description For `Cuda` and `HIP` backends, the `Kokkos::Impl::Graph` can be viewed as...
### Summary `Kokkos::kokkos_malloc` can be used to allocate memory in a memory space, allowing the memory to be tracked by `Kokkos` (among other benefits such as label). However, it seems...
### Summary The alias `host_mirror_type` defined in https://github.com/kokkos/kokkos/blob/e4cc6862c6ba7cd14d7a274660015a62efb62605/core/src/Kokkos_View.hpp#L657-L660 seems redundant with `HostMirror` https://github.com/kokkos/kokkos/blob/e4cc6862c6ba7cd14d7a274660015a62efb62605/core/src/Kokkos_View.hpp#L650-L655 The main difference is that one is using the `host_mirror_space`, the other creates a `Kokkos::Device`. Which alias...
**Describe the bug** It seems that it is currently impossible to take a `subview` of a rank-zero view, because `Kokkos::Impl::SubviewLegalArgsCompileTime` misses the required specialization. This is detrimental to generic code...
I'm using `ccache` `4.10`. The following compiler flag is unsupported: `-fdump-lang-switch-options=filename` (see *e.g.* https://gcc.gnu.org/onlinedocs/gcc/Developer-Options.html). Therefore, `ccache` will report my compile unit as *uncacheable*... It would be nice if it was...