Parallel graph algorithms are used to process large-scale graphs, such as social networks or search engine indexes. These algorithms typically use task-parallelism to decompose a graph into smaller sub-problems, and data-parallelism to process each sub-problem in parallel.