December 2017
Intermediate to advanced
718 pages
23h 14m
English
One of the most widely used parallel programming models today is MapReduce. MapReduce is easy both to learn and use, and is especially useful in analyzing large datasets. While it is not suitable for several classes of scientific computing operations that are better served by message-passing interface or OpenMP, such as numerical linear algebra or finite element and finite difference computations, MapReduce's utility in workflows frequently called “big data” has made it a mainstay in high performance computing. This chapter introduces the MapReduce programming model and the Hadoop open-source framework which supports it.
Read now
Unlock full access