EKAT

EKAT copied to clipboard

Potential problem with ExeSpaceUtils view_reduction and parallel_reduce

5

**Describe the bug** This was discovered when porting shoc_energy_integrals to small kernels. I was getting large differences in the outputs of the view_reductions when num_threads>1. I suspect the problem is...

jgfouca

Potential fix for Apple silicon build

3

For the past few months (~beginning of November), I haven't been able to build EKAT successfully on my mac laptop with an M1 chip that's on macos Monterey. First, the...

mjs271

**Is your feature request related to a problem? Please describe.** view_reduction currently handles just a single Pack as accumulator. The result is that in SHOC, there are two cases where...

ambrad

Do not make EKAT enable FPEs during initialize_ekat_session

4

I marked it as a bug, but it's more like a conceptual change. EKAT should not take care of enabling any FPEs when the ekat session is started. There are...

bartgol

Now that HIP support has been added (#225), and SYCL is going in soon (#230), we should probably start testing them as part of our PR testing. The problem is...

bartgol

Current kokkos submodule reference is not a tree

1

The version of EKAT that SCREAM uses has a Kokkos submodule reference that is not a tree. Github message: "This commit does not belong to any branch on this repository,...

ambrad

PR #218 added the ability to dispatch a kernel on host when CudaUVM is used, by creating a team policy on host. However, CudaUVM is still a device memory space,...

bartgol

See: https://github.com/E3SM-Project/scream/issues/1559 The WSM is throwing an error in debug for larger problems (ne30) when WSM is being used in a suboptimal way (taking and releasing individual slots instead of...

jgfouca

Allow specific log filenames to be supplied to `ekat::logging::Logger`

2

**Is your feature request related to a problem? Please describe.** Currently, all of the constructors for EKAT's `Logger` class append the MPI rank and `_logfile.txt` to the name provided. E3SM...

jeff-cohere

It is possible that the user tries to stress test some functionalities for very large arrays. In this case, we _might_ need 64 bits for their size. Note: I don't...

bartgol