| 摘要 |
PROBLEM TO BE SOLVED: To solve the problem in the conventional distribution processing on a network that the load of paralleling on a client PC is high because paralleling of a task is performed by the client PC and the other PCs being processor elements perform only processing and returning of the obtained paralleled task and that a network load is high because the PCs being the processor elements are required to process all the obtained tasks by themselves and assign task requests to more of the PCs as a result. SOLUTION: The PCs being the processor elements are provided with the client PC provided with a distribution processing calculation means for calculating the number of paralleling from the load of the task and its own operating situation, a parallel complier for paralleling the task, a hierarchical site describing means for adding its own address, a hierarchy and an identifier showing a site within the hierarchy to the paralleled task and a path storage means for storing a necessary time for obtaining an integrated result, the group of the addresses and a distribution ratio in a state that they are made to correspond to each other.
|
