| Autor: |
Baisheng Yang, Xiaomin Qiang, Yanyan Li, Xuehui Shao, Yan Chen, Zongjing Gu, Peng Sun |
| Rok vydání: |
2020 |
| Předmět: |
|
| Zdroj: |
2020 IEEE MTT-S International Conference on Numerical Electromagnetic and Multiphysics Modeling and Optimization (NEMO). |
| Popis: |
The parallel method of moment is firstly studied based on many-core supercomputer platform with a peak computing speed of more than 100PFlops , which is the first supercomputer of China built by all use independent technology. For the matrix solution process with intensive computation and high CPU utilization, the matrix multiplication operation with high concurrency degree is executed at the slave with strong computing power by using the xMath mathematical library of "Sunway TaiHu Light" system. The parallel efficiency of solving matrix is up to 90% using one-way communication LU algorithm named CALU, shows better scalability than open ScaLAPACK mathematical library. The numerical results show that CALU algorithm has a strong advantage in solving method of moment equations based on domestically-made supercomputer platform, which greatly speeds up the process of solving matrix. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
