What do understand by the granularity of a parallel system?

Granularity refers to the quantity of computation complete in parallel relative to the size of the entire program. In parallel computing, granularity is a qualitative measure of the percentage of computation to communication.  Through granularity of the system, parallel- processing systems can be distributed into two set: fine-grain systems and coarse-grain systems.  In fine-grained systems parallel parts are relatively little and which means more frequent communication. Fine-grain processing's had short computation to communication ratio and requires elevated communication above your head. In coarse grained systems parallel parts are relatively large and which means less computation and more communication. If granularity is too superior it is possible that the overhead required for communications and synchronization between tasks takes longer than the computation. On the other side, in coarse-grain parallel systems, relatively huge amounts of computational work are completed. Coarse-grain processing's has a high computation to communication ratio and imply more chances for performance increase.