Improving Statistical Calculation via Hybrid Parallel Processing with Shared and Distributed Memory Based Parallelization
Project/Area Number |
24500344
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Research Field |
Statistical science
|
Research Institution | Oita University |
Principal Investigator |
|
Project Period (FY) |
2012-04-01 – 2015-03-31
|
Project Status |
Completed (Fiscal Year 2014)
|
Budget Amount *help |
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2014: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2013: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2012: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
|
Keywords | 統計科学 / 計算機統計 / 並列計算 / 共有メモリ型並列計算 / 分散メモリ型並列計算 / 計算機統計学 |
Outline of Final Research Achievements |
It is investigated to improve statistical calculation with parallel processing algorithms under an environment with a group of computers having multicore CPUs. In order to implementing parallel algorithms, characteristics of shared memory parallelism and distributed memory parallelism are taken into account and hybrid type processing with both approaches is devised to achieve overall efficiency of algorithms. First, general methods of task management are examined to parallelize Monte Carlo quadrature as a hybrid parallel processing and it is followed by investigating ways to parallelize specific algorithms of actual statistical analyses, such as probability variation model for discrete response data and exact inference of logistic regression analysis.
|
Report
(4 results)
Research Products
(14 results)