| Project/Area Number |
25870226
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
High performance computing
Computational science
|
|---|
| Research Institution | Tokyo Institute of Technology |
|---|
Principal Investigator |
ONODERA Naoyuki 東京工業大学, 学術国際情報センター, 助教 (50614484)
|
|---|
| Project Period (FY) |
2013-04-01 – 2015-03-31
|
| Project Status |
Completed (Fiscal Year 2014)
|
| Budget Amount *help |
¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
Fiscal Year 2014: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
Fiscal Year 2013: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
|
|---|
| Keywords | ハイパフォーマンス・コンピューティング / GPU / 格子ボルツマン法 / ラージエディ・シミュレーション / 流体構造連成 / 大規模並列計算 |
|---|
| Outline of Final Research Achievements |
Turbulent phenomena are one of the important topics in CFD. In order to understand the detail of airflows, it is necessary to carry out large-scale CFD simulations. We have developed a CFD code based on LBM (Lattice Boltzmann Method) for GPU supercomputers. We calculated a large-scale wind simulation for a 10km × 10km area in Tokyo. We used 4,032 GPUs for the computation with 10,080 × 10,240 × 512 mesh. By executing a large-scale wind simulation, we confirm large-scale structure of wind and detailed winds behind buildings. The code is written in CUDA and the GPU kernel function is well tuned to achieve high performance on TSUBAME supercomputer. By introducing the overlapping technique between the GPU-to-GPU communication and the GPU kernel computation, we achieved 1.14 PFLOPS using 3968 GPUs in single precision on TSUBAME 2.5.
|
