2011 Fiscal Year Final Research Report
Skeletal Parallel Programming for Non-numerical Problems
Project/Area Number |
22800007
|
Research Category |
Grant-in-Aid for Research Activity Start-up
|
Allocation Type | Single-year Grants |
Research Field |
Software
|
Research Institution | The University of Tokyo |
Principal Investigator |
EMOTO Kento 東京大学, 大学院・情報理工学系研究科, 助教 (00587470)
|
Project Period (FY) |
2010 – 2011
|
Keywords | 並列プログラミング / 生成検査集約法 / 非数値計算問題 / 組み合わせ問題 / プログラム導出 / 半環 / スケルトン / 動的計画法 |
Research Abstract |
We have developed(1) a theory that provides an automatic optimization mechanism to derive efficient parallel programs from user programs naively-written in a concise notation for non-numerical problems on sequences, and(2) a library that provides the notation and the automatic optimization. This research has applied the skeletal parallel programming to non-numerical programs : The optimization has been developed based on the theory of skeletal parallel programming, and in our proposing GTA(Generate, Test, and Aggregate) programming a program is built by three kinds of components, each of which can be seen as a kind of skeletons. In addition, our GTA programming can be seen as a new programming framework for dynamic programming algorithms, in a point of view of program development/derivation.
|