Declarative Data Collections for Portable Parallel Performance based on oneAPI

Abstract

Collection Skeletons provide a novel, declarative approach to data collections through properties. They also help shield the app developer from parallel implementation details either by encapsulating implicit parallelism, or through explicit properties that match with parallel algorithmic skeletons. Implicit parallelism is implemented through providing concurrent data structures with member functions based on oneTBB; Explicit parallelism comes from parallel algorithmic skeletons based on SYCL.