Classification Based On Grain Size
This classification is based on identifying the parallelism in a program to be implemented on a multiprocessor system. The plan is to recognize the instructions in a program or sub-tasks that can be implemented in parallel. For instance, there are three declaration in a program and statements S1 and S2 can be swaped. It means, these are not chronological as shown in Figure 15. Then S1 and S2 can be implemented in parallel.
But it is not enough to check for the parallelism among processes or statements in a program. The conclusion of parallelism also depends on the following factors: