| Budget Amount *help |
¥43,420,000 (Direct Cost: ¥33,400,000、Indirect Cost: ¥10,020,000)
Fiscal Year 2019: ¥5,200,000 (Direct Cost: ¥4,000,000、Indirect Cost: ¥1,200,000)
Fiscal Year 2018: ¥5,200,000 (Direct Cost: ¥4,000,000、Indirect Cost: ¥1,200,000)
Fiscal Year 2017: ¥5,200,000 (Direct Cost: ¥4,000,000、Indirect Cost: ¥1,200,000)
Fiscal Year 2016: ¥13,650,000 (Direct Cost: ¥10,500,000、Indirect Cost: ¥3,150,000)
Fiscal Year 2015: ¥14,170,000 (Direct Cost: ¥10,900,000、Indirect Cost: ¥3,270,000)
|
|---|
| Outline of Final Research Achievements |
Given the specification of simulation as a large irregular graph, we have established a technique to efficiently and automatically generate a parallel high-performance simulation program.
Specifically, we firstly improved the parallelization rate of automatic program generation by dividing the graphs at a higher level of abstraction, which resulted in a 10 times speedup compared to fine-grained graph division. We secondly optimized the generated program according to a variety of parallel computers - the GPU, CPU clusters, and vector supercomputers -, which achieved an about two times speedup at most. Using the biological simulator that is one of the main applications of this work, users can get the benefits of various parallel computers without creating programs.
|
