An Algorithm Module Integration Framework (AMIF) for Computation Application
Many science applications need to deal with a lot of complex processing which requires coupling algorithm modules from multiple disciplines, and in turn demands advanced techniques to integrate these modules together effectively so that the data flows seamlessly. The role defined for the algorithm modules is to receive input data, calculate determined outputs and send the results back to the host applications. In this paper, we overview some processing applications required for such integration, and highlight the benefit for using an integration framework in these applications. Then we will introduce a reusable integration framework to isolate the host applications from the inherent complexity due to the nature of data handling, module integration and processing implementation sharing. Released from the time required to integrate the different processing modules, the engineer could pay more attention on defining the overall computation logic and more importantly the science content of the algorithm. Through examples, we demonstrate the feasibility of such integration framework and explore its possible use into a numerical computation application.
AMIF Computation Algorithm Modules 1Integration Framework
Qingqing Sun Qiang Liu Yang Liu
the School of Software, Tsinghua University, China the Schlumberger Beijing GeoScience Center, China
国际会议
武汉
英文
2007-09-21（万方平台首次上网日期，不代表论文的发表时间）