Abstract
AbstractIn parallel program partitioning, a program is split into tasks that can execute concurrently on a multiprocessor. In this chapter, we develop methods for statically partitioning certain types of parallel loops. Static program partitioning is attractive since the program partition is specified during the compilation phase, thereby eliminating one source of run-time overhead. Furthermore, static partitionings reduce communication overhead relative to dynamic schemes that attempt to improve processor utilization. Our partitioning algorithm is adaptive data partitioning (ADP), which chooses a partition based on data access patterns.KeywordsIteration SpaceCache LineRegular HexagonNeighborhood CommunicationCode SegmentThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
