Beschreibung

Presentation of a framework that provides OpenCL compatibility by implementing a memory model based on shared memory between host and compute device, which removes the overhead imposed by data transferences at global memory level, and local memories inside each accelerator.

Links und Ressourcen

Tags