Efficient compile-time/run-time contraction of fine grain data parallel codes

Threaded Runtime Support for Execution of Fine Grain Parallel Code on Coarse Grain Multiprocessors

“…This approach avoids the use of a separate thread for scheduling. Heuristics that schedule based on communication patterns were introduced by Neves and Schnabel in [17] and were shown to reduce the total number of context switches and improve application data locality. Briefly, our scheduling mechanism works by first attempting to switch to the last VP data were received from.…”