الفهرس | Only 14 pages are availabe for public view |
Abstract A token based scheduler is developed to enable efficient utilization of graphics processing unit in big-data clusters specifically for computer vision applications. The scheduler addresses the racing conditions that occur on the graphics processing unit due to simultaneous access by the parallel instances of the running application. The presented scheduler enables the porting of computer vision applications to big-data cluster with heterogeneous computing capabilities where multi-core central processing units exist alongside graphical processing unit |