Study on Acceleration of Three-Dimensional Method of Characteristics by GPU
-
摘要: 三维特征线方法可以精确求解任意几何堆芯的稳态多群中子输运方程,但同时也具有收敛慢、计算时间长的不足,需要研究相应的加速手段。图形处理器(GPU)计算由于具有速度快,能耗低的优点,被认为是未来高性能计算发展的方向之一。研究GPU计算加速三维特征线方法,并将其应用到三维特征线程序TCM中。借助统一计算设备架构(CUDA)的GPU计算,中央处理器(CPU)负责内存分配、有效增殖系数keff和源分布计算等逻辑性强或归约计算的处理,GPU执行特征线射线扫描细网求解细网通量。计算结果表明,经改写后的程序具有良好的加速效果。Abstract: The three-dimensional method of characteristics(MOC) can solve neutron transport equation for arbitrary geometry accurately.However,the MOC has some drawbacks: the convergence speed is slow and very time consuming.As a result,the research of acceleration of MOC is carried out.Compared with the CPU computing,the GPU computing,which is one of the most promising high performance computing,can achieve higher computing speed but with lower cost.And the development of general computing on GPU can be simplified with CUDA.To reduce the computing time and increase the computing efficiency,the study of three-dimensional MOC is performed and applied to the three-dimensional MOC code TCM.The computing results confirm the excellent acceleration of the code running on GPU.
-
Key words:
- Neutron Transport /
- Method of characteristics /
- GPU /
- CUDA
-
计量
- 文章访问数: 15
- HTML全文浏览量: 5
- PDF下载量: 0
- 被引次数: 0