Using FG to Reduce the Effect of Latency in Parallel Programs Running on Clusters

We use parallel computing to get high performance, and one of the most significant obstacles to high-performance computing is latency. We also use parallel computing-particularly distributedmemory clusters-to operate on massive datasets, taking advantage of the memory and disk space across the multiple nodes of a cluster. But working with massive data on a cluster induces latency in two forms: accessing data on disk and interprocessor communication.