Workload Allocation



next up previous contents index
Next: Data Decomposition Up: Basic Programming Techniques Previous: Tree Computations

Workload Allocation

In the preceding section, we discussed the common parallel programming paradigms with respect to process structure, and we outlined representative examples in the context of the PVM system. In this section we address the issue of workload allocation, subsequent to establishing process structure, and describe some common paradigms that are used in distributed-memory parallel computing. Two general methodologies are commonly used. The first, termed data decomposition or partitioning, assumes that the overall problem involves applying computational operations or transformations on one or more data structures and, further, that these data structures may be divided and operated upon. The second, called function decomposition, divides the work based on different operations or functions. In a sense, the PVM computing model supports both function decomposition (fundamentally different tasks perform different operations) and data decomposition (identical tasks operate on different portions of the data).