A fully-customized dataflow engine for 3D earthquake simulation with a complex topography
A fully-customized dataflow engine for 3D earthquake simulation with a complex topography
复制标题
用于复杂地形 3D 地震模拟的完全定制数据流引擎
DOI：
10.1007/s11432-020-2976-5
复制
发表时间：
2021-11
期刊：
影响因子：
--
通讯作者：
Guangwen Yang
中科院分区：
文献类型：
--
作者：
Bingwei Chen;Haohuan Fu;Wayne Luk;Guangwen Yang
随着HPC（高性能计算）发展到亿亿次时代，计算性能和功率效率的改进变得越来越重要。基于我们之前在Sunway TaihuLight上实现大规模地震模拟的工作，我们进一步探索通过可重构FPGA（现场可编程门阵列）设备上的完全定制硬件设计来改进应用的其他可能性。我们调查的可行性和一个完整的定点设计的潜在好处。我们首先进行了一个基于粗分辨率的模拟，以分析捕获总能量和能量分布的变量在空间和时间上所需的表示范围和精度。然后，我们得到一个完整的定点设计，确定合适的位宽的主要类别的变量，并动态地表示通过动态缩放方案的范围。最后，我们使用优化的定点设计运行的汶川地震的情况下，证明了在FPGA设备上支持大规模的科学模拟的潜力。结果表明，18位定点设计已经提供了汶川场景中几乎相同的地震事件描述，直到单精度浮点版本，并提供相当于13.1个英特尔至强金牌6154 18核CPU或2.10个Sunway 260核处理器的可持续性能，与Intel Xeon Gold 6154 18核CPU和Sunway 260核处理器相比，性能功耗比（能效）分别提高了15.3倍和3.72倍。
With HPC (high performance computing) evolving into the exascale era, improvements in computing performance and power efficiency have become increasingly more important. Based on our previous work on enabling earthquake simulations on a large scale on Sunway TaihuLight, we further explore other possibilities to improve the application through a fully-customized hardware design on reconfigurable FPGA (field programmable gate array) devices. We investigate the feasibility and the potential benefits of a complete fixed-point design. We first perform a coarse-resolution-based simulation to analyze the representation range and precision needed to capture both the total energy and the energy distribution of variables over space and time. We then derive a complete fixed-point design that identifies the suitable bitwidth for major categories of variables and dynamically represents the range through a dynamic scaling scheme. Finally, we use the optimized fixed-point design to run a case of the Wenchuan earthquake to demonstrate the potential of supporting large-scale scientific simulations on FPGA devices. The results demonstrate that an 18-bit fixed-point design already provides an almost identical description of the seismic events in the Wenchuan scenario down to a single-precision floating-point version and provides sustainable performance equivalent to 13.1 Intel Xeon Gold 6154 18-core CPUs or 2.10 Sunway 260-core processors, with performance per watt (power efficiency) improved by 15.3 and 3.72 times compared with the Intel Xeon Gold 6154 18-core CPUs and the Sunway 260-core processors, respectively.
登录
查看更多内容
DOI：
10.1109/sc.2012.106
发表时间：
2012-11
期刊：
2012 International Conference for High Performance Computing, Networking, Storage and Analysis
影响因子：
--
作者：
S. Habib;V. Morozov;H. Finkel;A. Pope;K. Heitmann;Kalyan Kumaran;T. Peterka;J. Insley;David Dani
通讯作者：
S. Habib;V. Morozov;H. Finkel;A. Pope;K. Heitmann;Kalyan Kumaran;T. Peterka;J. Insley;David Dani
DOI：
10.1007/978-1-4614-1791-0_10
发表时间：
2013
影响因子：
1
作者：
V. Medeiros;A. C. Barros;A. Silva-Filho;M. Lima
通讯作者：
V. Medeiros;A. C. Barros;A. Silva-Filho;M. Lima
影响因子：
5.8
作者：
A. Draut;P. Clift
通讯作者：
A. Draut;P. Clift
影响因子：
2.8
作者：
Zhang, Wei;Chen, Xiaofei
通讯作者：
Chen, Xiaofei
DOI：
10.1109/sc.2018.00043
发表时间：
2018-11
期刊：
SC18: International Conference for High Performance Computing, Networking, Storage and Analysis
影响因子：
--
作者：
Bingwei Chen;H. Fu;Yanwen Wei;Conghui He;Wenqiang Zhang;Yuxuan Li;W. Wan;Wei Zhang-;L. Gan;Zhenguo Zhang;Guangwen Yang;Xiaofei Chen
通讯作者：
Bingwei Chen;H. Fu;Yanwen Wei;Conghui He;Wenqiang Zhang;Yuxuan Li;W. Wan;Wei Zhang-;L. Gan;Zhenguo Zhang;Guangwen Yang;Xiaofei Chen