MCFX: a new parallel programming framework for multicore systems. Parallel computing is going mainstream with the now ubiquitous multicore and cluster systems. Despite the availability of many parallel programming models, there is no one-size-fits-all solution. Many application experts are still desperately looking for a better solution to meet their needs. We present a new parallel programming framework called MCFX to improve the productivity of shared memory parallel programming with a high-level parallel execution model. In MCFX, we introduce the concept of domain-specific parallel schedulers at application level. These schedulers are called executable containers. An executable object is a programming object associated with data and actions (processes). A MCFX executable container can concurrently execute the processes of the executable objects stored in the container. One can customize and instantiate many executable containers in a program. An important goal of project MCFX is to let the domain experts with basic knowledge of MCFX to design reusable parallel programming templates for their domains.
References in zbMATH (referenced in 1 article )
Showing result 1 of 1.
- Chen, Liang; Bairagi, Deepankar; Lin, Yuan: MCFX: a new parallel programming framework for multicore systems (2009)