Parallel Processing of CFD in the field of building and urban environmental science
| Project/Area Number |
04505003
|
|---|
| Research Category |
Grant-in-Aid for Developmental Scientific Research (A)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
建築環境・環境工学
|
|---|
| Research Institution | Institute of Industrial Science. University of Tokyo |
|---|
Principal Investigator |
MURAKAMI Syuzo Institute of Industrial Science, University of Tokyo Professor, 生産技術研究所, 教授 (40013180)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
MIZUTANI Kunio Sanken Setsubi Kogyo CO., LTD.Chief, 技術本部, 研究主査
UTUMI Yasuo Institute of Industrial Science, University of Tokyo Associate Professor, 助教授 (30168728)
MOTIDA Akasi Institute of Industrial Science, University of Tokyo Lecturer, 生産技術研究所, 講師 (00183658)
KATO Sinsuke Institute of Industrial Science, University of Tokyo Associate Professor, 生産技術研究所, 助教授 (00142240)
|
|---|
| Project Period (FY) |
1992 – 1993
|
| Project Status |
Completed (Fiscal Year 1993)
|
| Budget Amount *help |
¥47,000,000 (Direct Cost: ¥47,000,000)
Fiscal Year 1993: ¥6,000,000 (Direct Cost: ¥6,000,000)
Fiscal Year 1992: ¥41,000,000 (Direct Cost: ¥41,000,000)
|
|---|
| Keywords | numerical simulation / pararell processing / environmental engineering / pararell procedure / communication time / speedup ratio / SIMPLED method / algorithm / 並列計算 / 計算流体力学 / LES / 建築・都市環境 / 室内気流 / 建物周辺気流 |
|---|
| Research Abstract |
A computational procedure has been developed for simulating 2D and 3D incompressible steady state flows in enclosed spaces with parallel processing based an HX network computing system. The numerical procedure is based on a modified SIMPLE algorithm and has been parallelized with matrix calculation. Performance of the parallel computation has been examined through simulating laminar flows in enclosed spaces. In the case of computation with 256 processors, inter - processor communication tim ee occupys about 10% of total calculation (CPU executing) time. When the data structure (i. e. dimension and size of data matrix) of the numerical procedure matches the characteristics of the computer system, the speedup ratio of parallel computation is proportional to the number of processors.
|
Report
(3 results)
Research Products
(21 results)
