The current OpenMP spec has GPU offload features specifically for what was expected of the Sierra supercomputer. I'm not sure how relevant a paper that old (re

The current OpenMP spec has GPU offload features specifically for what was expected of the Sierra supercomputer. I'm not sure how relevant a paper that old (relatively, I hasten to add) is.