Alex Dutka

Results 60 comments of Alex Dutka

No, not yet @ajpowelsnl. As I mentioned in my first message, it is more an annoyance than a real problem and it got overtaken by more urgent matters in the...

Behavior should be exactly the same, to the notable exception of the value of ``source_loaded`` in the case of a wrong file extension. Indeed, with the current implementation ``source_loaded ==...

@maartenarnst Yes this is an unavoidable data structure when general unstructured meshes are involved, and I am convinced that having such a data structure would prove useful for solving other...

> Looks good, with the possible caveat that the cited paper is for Kokkos-3. @ajpowelsnl Is there a Kokkos4 paper? (either already published or soon to be)

I believe there was a significant performance overhead even with "simple" ``parallel_for`` loops, which constituted about 95% of the codebase at the time. Is it something that was implemented recently...

From an exterior perspective (i.e. feel free to disregard my comment since I don't know what I'm talking about 😉), it seems rather strange that this check is implemented Npatterns...

It seems like something went wrong here, but I won't try to fix it without knowing first if this is something you are interested in...

Seems like the problem has been fixed...

> The first thing (containers with overloaded math operators) is not something I really want inside Kokkos Core. The whole ISO C++ valarray issue is just too off-putting. Understandable. The...