Francesco Rizzi
Francesco Rizzi
`Kokkos_ENABLE_DEBUG_BOUNDS_CHECK` catches this , i opened https://github.com/kokkos/kokkos/pull/6578 to work on a more useful message
@ajpowelsnl thanks!
> `LayoutStride` from a subview of a higher-rank View might actually be reasonable to support. It would require more implementation genericity to handle it, though, since the internal view that...
Yeah Phil left last year and currently I/nga have no contract.. no one is so that I am aware of, so if you pick this up it would be great!
I have to refresh my memory on this but I will take a look. Maybe @PhilMiller can chime in if he has some quick guidance
hi @srinivasyadav18 , pinging @mzuzek since he is the one who posted #218
> I'll leave this up to @crtrott , who is more familiar with the customization design. Here's what I wrote on #218. > > > I do not like that...
> Basically the infinite recursion should only happen if the specialization (i.e. the Kokkos variant) isn't more special than the generic one which is completely unconstrained on the ExecPolicy see...
> @fnrizzi Just to clarify: Christian and I did some investigation, and we think the actual issue could be that your custom `matrix_vector_product` doesn't sufficiently constrain its parameters. yes, ok,...
@mhoemmen yes thanks for those links, very useful :) I will read them. Re "However, it wouldn't prevent other issues. For example, if users implementing a customization don't constrain its...