Unified schemes for directive-based GPU offloading

Directive-based GPU offloading using OpenACC or OpenMP target is a convenient way to port existing codes originally deveoped for multicore CPUs. Although OpenACC and OpenMP target provide similar features, both methods have pros and cons. OpenACC has better functions and an abundance of documents, but it is virtually for NVIDIA GPUs. OpenMP target supports NVIDIA/AMD/Intel GPUs but has fewer functions than OpenACC. In this talk, we will present a newly developed header-only library, Solomon (Simple Off-LOading Macros Orchestrating multiple Notations), to unify the interface for GPU offloading with the support of both OpenACC and OpenMP target.