|Schedule:||View OpenMPCon 2015 Program|
|Organisation:||Oak Ridge National Laboratory|
|Abstract:||The importance of computational accelerator technology for the future of high performance computing is widely acknowledged. Programmability, performance portability and developer productivity for these systems is a topic of growing importance.In this paper we examine two emerging specifications offering the promise of performance portability for codes targeting accelerator-based computing systems: OpenMP 4.0 and OpenACC 2.0.We show how these two standards compare, both in similarities and differences.
We present code comparisons to show how each API is used to parallelize representative code fragments. Furthermore, we give guidelines for developers wishing to convert codes from OpenACC 2.0 to OpenMP 4.0.