Francesco Rizzi
Francesco Rizzi
@ndellingwood https://github.com/kokkos/kokkos/pull/6629 fixes it for me when i try on dev-2
Can you confirm it works? Just to be safe :)
thanks @gregvw for pointing this out! https://github.com/kokkos/kokkos/pull/6645 fixes this
@brian-kelley one reason we have not yet added iota (see #4084) is exactly your point of not being easily parallelizable. We do have transform implemented so @mhoemmen 's idea can...
Should we provide something like `Kokkos::Experimental::fill_with_range` or `Kokkos::Experimental::fill_with_consecutive_range` or something else? I can take this on and add it to the algorithms. Or whatever variant we decide. I can work...
My suggestion below is purely a result of providing something that serves the usecase, and avoids calling it iota. I personally think that a more expressive name is better. ```cpp...
@crtrott @dalg24 @nliber @brian-kelley any thoughts on my proposal?
@Yaraslaut I just started looking at this, thanks for the PR ! 1. first comment: ```cpp auto kokkos_complex = Kokkos::complex{2.0,2.0}; cyl_bessel_j(2,kokkos_complex); ``` I think it would be useful to allow...
making this a draft following the comment: https://github.com/kokkos/kokkos/pull/5897#discussion_r1378422127
> > Search for an existing implementation led me to discover implementation of Bessel function by Amos[1]. You can see this is a Fortran code and i am not sure...