Daniel Arndt
Fixes #7086.
As discussed in the meeting today, we want to test with various C++ standard versions in the GitHub CI. This pull request tests with both C++17 and C++20 in the...
Enabling the `Serial` and `OpenMP` backend in a build using `-fsanitize=undefined` with `clang-18.1.5` reports ``` /tmp/kokkos/containers/unit_tests/TestVector.hpp:210:36: runtime error: signed integer overflow: 18647684 * -6116 cannot be represented in type 'int'...
Currently, we are always using `cudaMallocAsync` for allocating memory if the compiler is recent enough (and the user doesn't forcefully disable it). As pointed out in https://github.com/kokkos/kokkos-core-wiki/pull/525#discussion_r1615644650, this is not...
Our CI for `linux debug cuda-11` is failing when we use `CMake 3.29.0` since that `CMake` version declares more targets then `Kokkos` exports. This issue is tracked at https://github.com/kokkos/kokkos/pull/6898. The...
Alternative to #1118. Uses the same approach as in https://github.com/kokkos/kokkos/pull/7112 to avoid errors such as ``` --2024-06-28 16:01:41-- (try: 2) https://registrationcenter-download.intel.com/akdlm/irc_nas/19133/l_oneDPL_p_2022.0.0.25335.sh Connecting to registrationcenter-download.intel.com (registrationcenter-download.intel.com)|23.211.176.50|:443... connected. HTTP request sent, awaiting...
We need to add `-fsycl` as a linker flag when `SYCL` is used which we assume to be possibly true when using `IntelLLVM`.
I need the first commit adding some `RandomAccessIteror` member functions for `MdspanRandomAccessIterator` to compile the tests. The second commit fixes an apparent oversight in not trying to check the result...
# Description Without this fix NVML can't be found via standard CMake variables. With this, I can do ``` cmake -DVARIORUM_WITH_NVIDIA_GPU=ON -DNVML_ROOT=/usr/local/cuda-11.7/targets/x86_64-linux/ ../src ``` do find `NVML`. Requiring finding the...
Merge after #559 # Description Corresponds to #559 for Intel GPUs using APMIDG. Outputs look like ``` > ./variorum-print-verbose-energy-example _INTEL_GPU_ENERGY_USAGE Host: x1921c0s2b0n0, Socket: 0, DeviceID: 0, Energy: 682460.802429 J _INTEL_GPU_ENERGY_USAGE...