Data-driven methods for computer simulations are blooming in many scientific areas. The traditional approach to simulating physical behaviors relies on solving partial differential equations (PDEs). Since calculating these iterative equations is highly both computationally demanding and time-consuming, data-driven methods leverage artificial intelligence (AI) techniques to alleviate that workload. Data-driven methods have to be trained in advance to provide their subsequent fast predictions; however, the cost of the training stage is non-negligible. This article presents a predictive model for inferencing future states of a specific fluid simulation that serves as a use case for evaluating different training alternatives. Particularly, this study compares the performance of only CPU, multi-GPU, and distributed approaches for training a time series forecasting deep learning model. With some slight code adaptations, results show and compare, in different implementations, the benefits of distributed GPU-enabled training for predicting high-accuracy states in a fraction of the time needed by the computational fluid dynamics solver.
AbadiMBarhamPChenJ, et al. (2016) TensorFlow: a system for large-scale machine learning. In: Proceedings of the 12th USENIX Conference on Operating Systems Design and Implementation, OSDI’16. Savannah, GA, 2-4 November 2016, pp. 265–283.
2.
ArchibaldRChowED’AzevedoE, et al. (2020) Integrating Deep Learning in Domain Sciences at Exascale. In: NicholsJVerasteguiBMaccabeAB, et al. (eds), Driving scientific and engineering discoveries through the convergence of HPC, big data and AI. Cham: Springer International Publishing, pp. 35–50.
3.
AwanAABedorfJChuCH, et al. (2019) Scalable distributed dnn training using tensorflow and cuda-aware mpi: Characterization, designs, and performance evaluation. In: 2019 19th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID), Larnaca, Cyprus, 14-17 May 2019. pp. 498–507. DOI: 10.1109/CCGRID.2019.00064.
4.
BerkoozGHolmesPLumleyJL (1993) The proper orthogonal decomposition in the analysis of turbulent flows. Annual Review of Fluid Mechanics25(1): 539–575. DOI: 10.1146/annurev.fl.25.010193.002543.
5.
BhattDZhangZhangBWZuckermanZuckermanDM (2010) Steady-state simulations using weighted ensemble path sampling. The Journal of Chemical Physics133(1): 014110.
6.
BruntonSLNoackBRKoumoutsakosP (2020) Machine learning for fluid mechanics. Annual Review of Fluid Mechanics52(1): 477–508.
7.
ClimentJBasieroLMartínez-CuencaR, et al. (2018) Biological reactor retrofitting using CFD-ASM modelling. Chemical Engineering Journal348: 1–14.
8.
HochreiterSSchmidhuberJ (1997) Long short-term memory. Neural Computation9: 1735–1780. DOI: 10.1162/neco.1997.9.8.1735
9.
IserteSCarratalaPArnauR, et al. (2021) Modeling of wastewater treatment processes with hydroSludge. Water Environment Research93(12): 1–38.
10.
IserteSMacíasAMartínez-CuencaR, et al. (2022) Accelerating urban scale simulations leveraging local spatial 3D structure. Journal of Computational Science62: 101741. DOI: 10.1016/j.jocs.2022.101741.
11.
JouppiNPYoungCPatilN, et al. (2018) A domain-specific architecture for deep neural networks. Communications of the ACM61(9): 50–59. DOI: 10.1145/3154484
12.
KimBAzevedoVCThuereyN, et al. (2019) Deep fluids: a generative network for parameterized fluid simulations. Computer Graphics Forum38(2): 59–70.
13.
KingmaDPBaJL (2015) Adam: a method for stochastic optimization. In: 3rd International conference on learning representations, ICLR 2015 - conference track proceedings. San Diego, CA, 7-9 May 2015.
14.
MaasALHannunAYNgAY (2013) Rectifier nonlinearities improve neural network acoustic models ICML workshop on deep learning for audio. Speech and Language Processing28.
15.
MalikALuMWangN, et al. (2018) Detailed performance analysis of distributed tensorflow on a gpu cluster using deep learning algorithms. In: 2018 New York Scientific Data Summit (NYSDS). New York, NY, 06-08 August 2018. pp. 1–8.
Obiols-SalesOVishnuAMalayaN, et al. (2020) CFDNet. In: Proceedings of the 34th ACM International Conference on Supercomputing, New York, NY, 29 June 2020-2 July 2020. DOI: 10.1145/3392717.3392772
18.
PaulAMozaffarMYangZ, et al. (2019) A real-time iterative machine learning approach for temperature profile prediction in additive manufacturing processes. In: Proceedings - 2019 IEEE International Conference on Data Science and Advanced Analytics, Washington, DC, 5-8 October 2019. pp. 541–550.
19.
Quang-HungNDoanHThoaiN (2020) Performance evaluation of distributed training in tensorflow 2 In: 2020 International conference on advanced computing and applications (ACOMP), Quy Nhon, Vietnam, 25-27 November 2020. pp. 155–159. DOI: 10.1109/ACOMP50827.2020.00031
20.
Ramirez-GargalloGGarcia-GasullaMMantovaniF (2019) Tensorflow on state-of-the-art hpc clusters: A machine learning use case, 2019 19th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGRID), Larnaca, Cyprus, 14-17 may 2019. pp. 526–533. DOI:10.1109/CCGRID.2019.00067
21.
RojekKWyrzykowskiRGepnerP (2021) AI-accelerated CFD simulation based on OpenFOAM and CPU/GPU computing computational science – ICCS 2021. Berlin, Germany: Springer International Publishing, pp. 373–385.
22.
RościszewskiPKrzywaniakAIserteS, et al. (2022) Adaptation of AI-accelerated CFD Simulations to the IPU PlatformProceedings of the 1st Workshop on Applications of Machine Learning and Artificial Intelligence in High-Performance Computing. WAML-HPC), p. PPAM’22. ISBN in press.
23.
SergeevABalsoMD (2018) Horovod: fast and easy distributed deep learning in tensorFlow. Available at: http://arxiv.org/abs/1802.05799
24.
SrivastavaSDamodaranMKhooBC (2019) Machine learning surrogates for predicting response of an aero-structural-sloshing system. ArXiv Preprint arXiv:1911.10043.
25.
TompsonJSchlachterKSprechmannP, et al. (2017) Accelerating eulerian fluid simulation with convolutional networks. In: Proceedings of the 34th International Conference on Machine Learning - Volume 70, Sydney, 06-11 August 2017. p. 3424–3433.
26.
VinuesaRBruntonSL (2021) The potential of machine learning to enhance computational fluid dynamics. arXiv preprint arXiv:2110.02085.
27.
XiaoDHeaneyCEMottetL, et al. (2019) A reduced order model for turbulent flows in the urban environment using machine learning. Building and Environment148: 323–337.
