Automatic SMT Threading for OpenMP Applications on the Intel Xeon Phi Co-processor

Simultaneous multithreading is a technique that can im-prove performance when running parallel applications on the Intel Xeon Phi co-processor. Selecting the most effi-cient thread count is however non-trivial, as the potential increase in efficiency has to be balanced against other, po-tentially negative factors such as inter-thread competition for cache capacity and increased synchronization overheads. In this paper, we extend CRUST (ClusteR-aware Under-subscribed Scheduling of Threads), a technique for finding the optimum thread count of OpenMP applications run-ning on clustered cache architectures, to take the behavior of simultaneous multithreading on the Xeon Phi into ac-count. CRUST can automatically find the optimum thread count at sub-application granularity by exploiting applica-tion phase behavior at OpenMP parallel section boundaries, and uses hardware performance counter information to gain insight into the application’s behavior. We implement a CRUST prototype inside the Intel OpenMP runtime library and show its efficiency running on real Xeon Phi hardware