Christian Trott

icon company Sandia National Laboratories icon location Albuquerque, NM, USA icon location High Performance Computing developer with extensive experience in modern architectures (APUs, GPUs, MIC). Lead developer of the @kokkos programming model.

Results 519 comments of Christian Trott

Define new type trait to identify cuda-type execution space in deep copy

Hey I gonna close this for now, trying to do a little bit of cleanup. You can reopen if and when you get back to it.

Ensure kernels submitted by multiple threads to synchronous execution spaces are enqueued correctly

We need to look at overhead for this, also its not clear to me how fence() would help with the is_running thing here?

Segfaults with AppleClang + --kokkos-num-threads>=4 on m1 mac

Do you still see this issue? We haven't been able yet to reproduce but could try again.

Determining a View's SharedAllocation size without actually creating and allocating memory for the view

Actually with our mdspan transition this will just work. Because the required_span_size is a property of the mapping.

Determining a View's SharedAllocation size without actually creating and allocating memory for the view

No this is part of our implementation backport to C++17

Initial BasicView/mdspan integration

A few more general thoughts: - `BasicView` (or whatever we will call it in the end) should be effectively `mdspan` with `Kokkos::View` owning semantics - it owns the tracking object...

Apps using Kokkos

Preliminary App list. We actually should have around 100 projects using Kokkos now - but we do not yet have a comprehensive list :-( | AppName | Area | Institution...

Apps using Kokkos

Thanks Greg for using the template ;-)

Apps using Kokkos

Hey Kendall can we break Sierra down a bit considering that it is a number of applications? Like we are not putting in Trilinos as one monolithic 5M line 30...

Basic parallel version of std::iota

One idea would be a `deep_copy` from range i.e. `deep_copy(v,pair{begin,end})`