Hello LLVM,
Openmp target offloading (GPUs, at present) has turned into a substantial project in it's own right with mostly orthogonal concerns to openmp host. There's been a multiyear effort to get nvidia and amdgpu targets running, with host + device + clang + llvm code implemented, weekly calls and annual gsoc students.
I would like to propose a code owner for openmp target offloading. Specifically I would like to nominate Johannes, as in my opinion he's been doing that role for the last couple of years anyway, and is a reliable choice for first point of contact for new contributors.
Objections? Support? Other?
Thanks all,
Jon