基于CPU/GPU异构系统架构的高超声速湍流直接数值模拟研究
党冠麟,刘世伟,胡晓东,张鉴,李新亮

Direct Numerical Simulation of Hypersonic Turbulence Based on CPU/GPU Heterogeneous System Architecture
Dang Guanlin,Liu Shiwei,Hu Xiaodong,Zhang Jian,Li Xinliang
表2 线程块大小与函数运算耗时
Table 2 Thread block size and the time of function operation
Block大小 时间(ms)
32*1*1 821.47
32*2*1 449.34
32*2*2 349.12
32*4*2 349.37
32*4*4 371.04