Sections of this module:Introduction
Case 1 - EPP
Case 2 - Data Sharing
Case 3 - On-demand Computing
Case 4 - Remote Access
Case 5 - Research App
Case 6 - Reproducible Research
Case 7 - Training
Case 8 - Big Data demands
View all sections on one page (Printer friendly)
Case 1 - EPP
You have an “embarrassingly parallel problem” (short: EPP). “Embarrassingly parallel problems” are characterized by being trivially parallelizable: no complex methods have to be applied to solve the problem by means of parallel methods, for example solving each part of the problem in a separate computer. This may save you huge amounts of time for your results to be finished.
For example, a very large data set can be chopped into pieces which are then dispatched to various computers for processing; when finished, the resulting data is re-assembled.
Or in another example, copies of a smaller data set are distributed across computers to perform different computations on it, and when all the individual analyses of the data are finished, the results are summarized.
The individual computers don’t have to be super fast, but instead the power lies in having a huge number of computers working at solving the problem simultaneously.
Software like MapReduce can be used to manage splitting the problem into several pieces and dispatching them to different computers.