Arturo Vargas

We are missing an example that exercises the kernel naming capability in RAJA::forall. At the moment I can't seem to compile the example provided in the RAJA docs (https://raja.readthedocs.io/en/develop/sphinx/user_guide/feature/reduction.html#experimental-reduction-interface).

Offer arbitrary global indexing for RAJA launch

1

For deeply nested loops consider an interference which allows for arbitrary global threading Potential idea: `Raja::GetThreadIndex< Dim >() ` ? User request: Yohann Dudouit

Users would like a simple interface to express the following types of computing patterns: ``` RAJA::Launch(..., [&]() { // Loop for z thread with inner loops for x and y...

This issue pulls in various reported issues to create a list of tasks. - [x] CHAI/Plugin support (https://github.com/LLNL/RAJA/issues/1379, https://github.com/LLNL/RAJA/issues/1333) : in progress https://github.com/LLNL/RAJA/pull/1381 - [ ] Add Sycl unit test...

We are missing support for OMP Target in RAJA::launch.

Extend RAJA launch docs

2

Further describe capabilities such as `cuda_global_thread_{x,y,z}`, tiling, omp parallel regions. Missing docs reported in https://github.com/LLNL/RAJA/issues/1057 - [x] Basic overview of API - [x] Run-time CPU/GPU switching - [x] Overview of...

Draft: Add SW4CK kernels

1

Adding SW4CK kernels 2 and 5 as they exhibit a performance gap between RAJA and "native variants". Source https://github.com/LLNL/SW4CK.

Address two bugs in SYCL launch backend 1. Ensures async boolean works properly 2. Eliminates memory leaks when using non-trivially copyable lambdas.

- [x] Add 3D version - [x] Further optimize with vendor linear algebra libraries - [ ] Clean up / organize

Reopened this PR in hoped that it would have better luck with CI.