This project describes a programming interface called PARRAY (or Parallelizing ARRAYs) that supports systemlevel succinct programming for heterogeneous parallel systems. PARRAY extends mainstream C and C++ programming with new array typed that contain additional information about the memory type, the layout of the elements in memory and the distribution of data over multiple memory devices. The users only need to learn a unified style of programming for all major parallel architectures including multicore (Pthread/OpenMP), clustering (MPI), distributed memory sharing (G lobal Arrays), GPU (CUDA) and manycore (MIC’s OpenMP Offload). The compiler will generate high-performance code according to the typing information contained in the source. This leads to shorter, more portable and maintainable parallel codes, while the programmer still has control over performance-related features necessary for performance optimization.
References in zbMATH (referenced in 1 article )
Showing result 1 of 1.