kokkos VS stdBLAS

I’m a novice user of Kokkos. Write code once for openmp, CUDA, and other parallel execution backends. It was designed with scientific computing applications in mind. Some numerics tools are implemented in “Kokkos kernels”, most of the BLAS operations are included iirc.

I would suggest using a C++ abstraction around thread parallelism. This will make your code easier to read and more concise, and will also make it easier to switch between different thread-parallel programming models. Kokkos is a lovely example of such an abstraction, but there are others. Modern C++ even has thread-parallel standard algorithms. Bryce Adelstein Lelbach's CppCon 2021 talk describes these.

It’s a work in progress, but Kokkos and the associated Kokkos Kernels are probably the closest thing to what you’re asking for.

Kokkos implements a programming model in C++ for writing performance portable applications targeting all major HPC platforms. It provides abstractions for both parallel execution of code and data management with a variety of backends including, but not limited to: CUDA, HIP, OpenMP, HPX, and Pthreads, with backends for OpenMPTarget and SYCL currently under development.

You can have a look at https://github.com/kokkos/stdBLAS (it's an implementation of the proposed linear algebra extensions for future C++).

It has the same BSD license as Kokkos: https://github.com/kokkos/stdBLAS/blob/main/LICENSE

The BLAS and LAPACK have C interfaces, and they generally don't allocate memory unless you ask them. (Some BLAS implementations allocate internal storage for rearranging matrix data.) There are a few C++ BLAS wrappers; here is one of them.

The second one is stdBLAS by kokkos. It doesn't even use cuBLAS but it's a reference implementation of P1673. This paper describe what could be a blas API in the std. I would recommend trying to do something equivalent on your side.