基于“嵩山”超级计算机系统的大规模管网仿真

doi:10.19678/j.issn.1000-3428.0063418

摘要/Abstract

摘要： 供水管网仿真广泛应用于城市供水输配调度，是城市供水管网监测与维护的重要技术手段。由于在面向城市级的大规模管网中产生了海量的计算数据，因此在一般计算平台上无法满足管网仿真计算的算力需求。为提升城市级供水管网仿真的计算效率，提出一种有效的并行化方案。基于“嵩山”超级计算机系统采用中央处理器+数据缓存单元（CPU+DCU）架构，利用其在密集数据计算方面的优势，对“嵩山”超级计算机进行供水管网仿真。参照可移植性异构计算接口（HIP）异构编程模型，在“嵩山”超级计算机上实现供水管网仿真的异构计算，并结合管道数据分割方案，使用消息传递接口开启多进程以实现DCU加速数据通信传递。通过重定义数据类型解决计算过程中结构体传输问题，实现单节点内多DCU的大规模密集计算。在不同计算平台和多种计算策略仿真上的对比结果表明，与传统x86平台相比，该优化方案在小规模数据与大规模数据上的加速比分别达到5.269、10.760，与采用计算统一设备架构异构编程模型的传统GPU异构平台相比，计算性能有明显提高。

关键词: 中央处理器+数据缓存单元架构, 数据缓存单元加速器, 仿真计算, 可移植性异构计算接口, 消息传递接口

Abstract: Water supply pipeline network simulation is widely performed for urban water supply transmission and distribution scheduling.Moreover, it is an important technical approach for monitoring and maintaining urban water supply networks. The amount of calculation data generated in the city-level large-scale pipeline network is significant;therefore, they are not available on general computing platforms.To satisfy computing power requirements for performing pipeline network calculations through simulation and to improve the efficiency of the calculations, this paper proposes an efficient parallelization scheme and uses the Central Processing Unit+Data Cache Unit(CPU+DCU) architecture based on the "Songshan" supercomputer system to exploit its advantages in the calculation of dense data to simulate the water supply network.Subsequently, based on the Heterogeneous-computing Interface for Portability(HIP) heterogeneous programming model, a heterogeneous calculation of the water supply pipe network is implemented numerically on the "Songshan" supercomputer.Using the existing pipeline data segmentation scheme, the Message Passing Interface(MPI) is used to launch multiple processes to allow the DCU to accelerate the transmission of data communication.The data type is defined, the structure transmission problem encountered during the calculation is solved, and the large-scale intensive calculation of multiple DCU in a single node is realized.A comparison between the simulation results of different computing platforms and multiple computing strategies show that the speedup of the optimization scheme and implementation method proposed herein is 5.269 and 10.760 for small-and large-scale data, respectively, which are better than that of the classical x86 platform.Additionally, by implementing Compute Unified Device Architecture(CUDA) heterogeneous programming, the computing performance achieved is better than that of a model using the classical GPU heterogeneous platform.

Key words: Central Processing Unit+Data Cache Unit(CPU+DCU) architecture, DCU accelerator, simulation calculation, Heterogeneous-computing Interface for Portability(HIP), Message Passing Interface(MPI)

中图分类号:

TP391

杨周凡, 韩林, 李冰洋, 谢景明, 韩璞, 刘勇杰. 基于“嵩山”超级计算机系统的大规模管网仿真[J]. 计算机工程, 2022, 48(9): 155-161.

YANG Zhoufan, HAN Lin, LI Bingyang, XIE Jingming, HAN Pu, LIU Yongjie. Large-Scale Pipeline Network Simulation Based on “Songshan” Supercomputer System[J]. Computer Engineering, 2022, 48(9): 155-161.

https://www.ecice06.com/CN/Y2022/V48/I9/155

图/表 11

20220924180456

20220924180500

20220924180504

20220924180508

20220924180513

20220924180517

20220924180520

20220924180524

20220924180528

20220924180532

20220924180535

参考文献

[1] WANG H, MENG H, ZHU T.New model for onsite heat loss state estimation of general district heating network with hourly measurements[J].Energy Conversion and Management, 2018, 157:71-85.
[2] WANG H, MENG H.Improved thermal transient modeling with new 3-order numerical solution for a district heating network with consideration of the pipe wall's thermal inertia[J].Energy, 2018, 160:171-183.
[3] FANG T T, KONTU K, LAHDELMA R.Estimation of the state of a district heating network based on consumption measurements[C]//Proceedings of the 21st International Conference on Power and Energy Systems and Applications.[S.1.]:ACTA Press, 2012:1356-1367.
[4] 黄凯锋, 刘文君.北京某水厂模拟管网内部腐蚀特性分析[J].中国给水排水, 2007, 23(13):63-65. HUANG K F, LIU W J.Analysis of internal corrosion characteristic of simulated pipe-loop systems in a waterworks in Beijing[J].China Water & Wastewater, 2007, 23(13):63-65.(in Chinese)
[5] WOLDEYOHANNES A D, MAJID M A A.Simulation model for natural gas transmission pipeline network system[J].Simulation Modelling Practice and Theory, 2011, 19(1):196-212.
[6] LAI J Q, YU H, TIAN Z Y, et al.Hybrid MPI and CUDA parallelization for CFD applications on multi-GPU HPC clusters[J].Scientific Programming, 2020, 20:1-15.
[7] LIU S, ZOU B, ZHANG L M, et al.Heterogeneous CPU GPU-accelerated FDTD for scattering problems with dynamic load balancing[J].IEEE Transactions on Antennas and Propagation, 2020, 68(9):6734-6742.
[8] SONG P T, ZHANG Z J, ZHANG Q, et al.Implementation of the CPU/GPU hybrid parallel method of characteristics neutron transport calculation using the heterogeneous cluster with dynamic workload assignment[J].Annals of Nuclear Energy, 2020, 135:106957.
[9] MENG W W, CHENG Y G, WU J Y, et al.GPU parallel acceleration of transient simulations of open channel and pipe combined flows[J].Earth and Environmental Science, 2019, 240:052025.
[10] ALLEC S I, SUN Y J, SUN J N, et al.Heterogeneous CPU+GPU-enabled simulations for DFTB molecular dynamics of large chemical and biological systems[J].Journal of Chemical Theory and Computation, 2019, 15(5):2807-2815.
[11] YANG S C, WANG Y L.A hybrid MPI-CUDA approach for nonequispaced discrete Fourier transformation[J].Computer Physics Communications, 2021, 258:107513.
[12] SABER H, MAZAHERI H, RANJBAR H, et al.Utilization of in-pipe hydropower renewable energy technology and energy storage systems in mountainous distribution networks[J].Renewable Energy, 2021, 172:789-801.
[13] WANG H, WANG H Y, ZHU T, et al.A novel model for steam transportation considering drainage loss in pipeline networks[J].Applied Energy, 2017, 188:178-189.
[14] WANG H, WANG H Y, HAIJIAN Z, et al.Optimization modeling for smart operation of multi-source district heating with distributed variable-speed pumps[J].Energy, 2017, 138:1247-1262.
[15] 郝正航, 张宏宇, 陈卓.一种天然气管网与电网耦合系统的实时仿真模型与系统:中国, 106707795A[P].2017-05-24. HAO Z H, ZHANG H Y, CHEN Z.Real-time simulation model and system of natural gas pipeline network and power grid coupling system:China, 106707795A[P].2017-05-24.(in Chinese)
[16] RUAN Y, ZHANG X L, CHEN J N.Optimization of pipe network simulation algorithm for tree-shaped water injection system in large-scale oilfield[J].International Journal of Pattern Recognition and Artificial Intelligence, 2021, 35(13):156-167.
[17] 刘青昆, 马名威, 杨荣杰, 等.基于MPI+CUDA环境的静电相互作用能并行求解[J].计算机应用与软件, 2012, 29(11):35-38. LIU Q K, MA M W, YANG R J, et al.Parallel solving electrostatic interaction energy based on MPI+CUDA environment[J].Computer Applications and Software, 2012, 29(11):35-38.(in Chinese)
[18] SHAFI A, HASHMI J M, SUBRAMONI H, et al.Efficient MPI-based communication for GPU-accelerated dask applications[C]//Proceedings of the 21st IEEE/ACM International Symposium on Cluster, Cloud and Internet Computing.Washington D.C., USA:IEEE Press, 2021:277-286.
[19] SNIR M.MPI:the complete reference[M].Cambridge, USA:MIT Press, 1996.
[20] JACOBSEN D, THIBAULT J, SENOCAK I.An MPI-CUDA implementation for massively parallel incompressible flow computations on multi-GPU clusters[C]//Proceedings of the 48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition.[S.1.]:AIAA Press, 2010:522.
[21] 梁党卫, 张永哲, 李克文, 等.子群分层的粗粒度粒子群优化算法[J].计算机工程与设计, 2019, 40(2):389-393. LIANG D W, ZHANG Y Z, LI K W, et al.Coarse-grained particle swarm optimization algorithm for subgroup stratification[J].Computer Engineering and Design, 2019, 40(2):389-393.(in Chinese)
[22] OYARZUN G, BORRELL R, GOROBETS A, et al.MPI-CUDA sparse matrix-vector multiplication for the conjugate gradient method with an approximate inverse preconditioner[J].Computers & Fluids, 2014, 92:244-252.
[23] GAO T, GUO Y F, ZHANG B Y, et al.Memory-efficient and skew-tolerant MapReduce over MPI for supercomputing systems[J].IEEE Transactions on Parallel and Distributed Systems, 2020, 31(12):2734-2748.

选择文件类型/文献管理软件名称

选择包含的内容