Digitalisierung
Refine
Year of publication
- 2007 (2) (remove)
Document Type
- Part of a Book (2) (remove)
Language
- English (2)
Has Fulltext
- no (2)
Is part of the Bibliography
- no (2)
Keywords
- GCM (1)
- Grid Programming (1)
- Higher-Order Components (HOCs) (1)
- Loop Parallelization (1)
- genome processing (1)
- higher-order components (1)
- middleware (1)
- sequence alignment (1)
Institute
This work integrates two distinct research areas of parallel and distributed computing, (1) automatic loop parallelization, and (2) component-based Grid programming. The latter includes technologies developed within CoreGRID for simplifying Grid programming: the Grid ComponentModel (GCM) and Higher- Order Components (HOCs). Components support developing applications on the Grid without taking all the technical details of the particular platform type into account (network communication, heterogeneity, etc.). The GCMenables a hierarchical composition of programpieces and HOCs enable the reuse of component code in the development of new applications by specifying application-specific operations in a program via code parameters. When a programmer is provided, e. g., with a compute farm HOC, only the independent worker tasks must be described. But, once an application exhibits data or control dependences, the trivial farm is no longer sufficient. Here, the power of loop parallelization tools, like LooPo, comes into play: by embedding LooPo into a HOC, we show that these two technologies in combination facilitate the automatic transformation of a sequential loop nest with complex dependences (supplied by the user as a HOC parameter) into an ordered task graph, which can be processed on the Grid in parallel. This technique can significantly simplify GCM-based systems which combine multiple HOCs and other components. We use an equation system solver based on the successive overrelaxation method (SOR) as our motivating application example and for performance experiments.
Computational grids combine computers in the Internet for distributed data processing and are an attractive platform for the data-intensive applications of bioinformatics. We present an extensible genome processing software for the grid and evaluate its performance. Our software was able to discover previously unknown circular permutations (CP) in the ProDom database containing more than 70MB of protein data. A specific feature of our software is its design as a component: the Alignment HOC, a Higher-Order Component that makes use of the latest Globus toolkit as grid middleware. Besides genome data, the Alignment HOC accepts plugin code for processing this data as its input, and contains all the required configuration to run the component on top of Globus, thus, freeing the non-grid-expert user from dealing with grid middleware. Instead of writing data distribution procedures and configuring the middleware appropriately for every new algorithm, Alignment HOC users reuse the existing component and only write application-specific plugins. To maintain plugins persistently in a reusable manner, we built a web-accessible plugin database with a comfortable administration GUI. The flexible component-based implementation makes it easy to study CPs in other databases (e.g. UniProt/Swiss-Prot) or to use an alignment algorithm different than the standard Needleman-Wunsch. For the efficient distribution of workload, we developed a library of group communication operations for HOCs.