A Fine-grained Parallel Programming Model for Grid Computing

In this paper, a new fine-grained parallel programming model---- Thin Kernel model is brought forth. In this model, the partitions of the parallel tasks are separated from the computational kernel of the problem. Thin Kernel model produces the parallel tasks dynamically at runtime when the tasks are being scheduled which makes the task assigning method in Thin Kernel model a real distributed one. Thin Kernel model uses dynamic classes loading and on-demand code transport. The collaboration of these technologies paves the way for large-scale parallel computing over the wide-area network. 1