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

`Kokkos::subview` does not compile for a rank-zero view: missing one template specialization for `Kokkos::Impl::SubviewLegalArgsCompileTime`

I think this will be covered by the mdspan transition soon (hopefully).

Sycl unit failures with intel/2024.1.0 on PVC arch

So are you saying it works with something older than 2024.1.0 and it works with 2024.2.0?

Introduce `Kokkos::TypeInfo<T>::name()` as a replacement for `typeid(T).name()`

I did some compile time check and this new thing wins over typeid by a mile: ```c++ #include #include template bool instantiate_one() { #ifdef TYPEID auto s = typeid(Kokkos::View).name(); return...

Introduce `Kokkos::TypeInfo<T>::name()` as a replacement for `typeid(T).name()`

NVCC 11.7 its 7.7s vs 1m37s

Memory Pool size setting for cudaMallocAsync

We may wanna organize a meeting for a larger discussion.

Memory Pool size setting for cudaMallocAsync

Closing: we anyway moved away again from cudaMallocAsync due to too much problems downstream since it doesn't allow IPC. We will revisit memory pools as an explicit feature.

update README

Add the cmake option.

Add pitched allocation example

I do not understand how this does what you want with respect to pitched allocations?

Add pitched allocation example

I think the accessor needs to store a pitch or some information that every K elements there are 4 extra bytes or whatever.

Add pitched allocation example

Also your argument with the alignment doesn't make much sense to me. The object can't say that it is 12byte aligned anyway for pitched allocations, since the whole purpose is...