Scientific Applications Sample Clauses

Scientific Applications. Scientific workloads typically involve huge quantities of data and very complex computations required to analyse the data and turn it into the knowledge according to the Computational Science Education Reference Desk (CSERD), see Figure 2-5. Moreover, the knowledge must be further organized, classified, and stored, to be accessed by researchers and practitioners, and to serve as a basis to more computations. Good examples are computational chemistry and bioinformatics where computations of human genome can easily require thousands of computing hours when sequentially computed on a modern computer, as well as thousands of Gigabytes of storage. These workloads traditionally require highly parallel infrastructure designs and low latency access to the vastly scalable storage resources. To achieve this, data integration and sharing solutions, such as data warehousing, allow researchers to tap into multiple heterogeneous data sources. Creation and maintenance of data warehouses with hundreds of data sources, as well as the technical effort and costs of the associated software are affordable to a limited number of very large organizations.