Research Progress on CPU/GPU Heterogeneous Parallelism and Efficient Solvers in Large-Scale Complex Reservoir Simulation
Kang Wang1, Xiang Rao1,2,3,*
1 School of Petroleum Engineering, Yangtze University, Jingzhou 434023, Hubei, China. 2 State Key Laboratory of Low Carbon Catalysis and Carbon Dioxide Utilization (Yangtze University), Wuhan 430100, China
3 Western Research Institute, Yangtze University, Karamay 834000, China.
This paper aims to systematically review the latest evolutionary routes of High-Performance Computing (HPC) in the field of reservoir numerical simulation. First, starting from the underlying hardware architecture, the article details the fundamental differences in data transmission mechanisms, programming complexity, and acceleration performance between the CPU/GPU hybrid offload mode and the full GPU native architecture. Second, it deeply analyzes key algorithmic breakthroughs for fractured reservoirs, focusing on the parallel assembly strategy for Non-Neighbor Connections (NNC) in the Embedded Discrete Fracture Model (EDFM) on GPUs, as well as multi-color DILU and CPR-AMG two-stage preconditioning technologies adapted for GPU many-core architectures. Third, based on industrial benchmark data from mainstream simulators such as tNavigator, Stone Ridge Echelon, and the open-source OPM Flow, the acceleration effectiveness and scalability of heterogeneous parallelism are quantitatively evaluated across models of varying scales. Finally, the paper critically points out deep-seated contradictions in current technologies regarding calculation accuracy on non-K-orthogonal grids, the VRAM capacity “ceiling” effect, and cross-platform portability, while looking forward to the convergent development trends of mixed-precision computing and physics-constrained AI preconditioning technologies.
Keywords: Large-scale reservoir simulation; GPU acceleration; Heterogeneous parallelism; Linear solvers; Embedded Discrete Fracture Model (EDFM); Non-Neighbor Connections (NNC)
Full Text
Download the full text PDF for viewing and using it according to the license of this paper.
Funding
This research was no funding provided.
References
Meng, X., He, X., Hu, C., & Lu, X. (2025). A review of parallel computing for large-scale reservoir numerical simulation. Archives of Computational Methods in Engineering.
Andersen, T. M., Torben, J., Lye, K. O., & Lie, K. A. (2025, March). A comparison of DILU and ILU(0) as GPU-accelerated preconditioners. SPE Reservoir Simulation Conference. Society of Petroleum Engineers.
Esler, K., Mukundakrishnan, K., Natoli, V., Shumway, J., Zhang, Y., & Gilman, J. (2014, September). Realizing the potential of GPUs for reservoir simulation. 14th European Conference on the Mathematics of Oil Recovery (ECMOR XIV). European Association of Geoscientists & Engineers.
Liu, H., Yu, S., Chen, Z., Hsieh, B., & Shao, L. (2012, April). Parallel preconditioners for reservoir simulation on GPU. SPE Latin America and Caribbean Petroleum Engineering Conference. Society of Petroleum Engineers.
Liu, H., Wang, K., Chen, Z., Jordan, K. E., Luo, J., & Deng, H. (2015, October). A parallel framework for reservoir simulators on distributed-memory supercomputers. SPE/IATMI Asia Pacific Oil & Gas Conference and Exhibition. Society of Petroleum Engineers.
Appleyard, J. R., Appleyard, J. D., Wakefield, M. A., & Desitter, A. L. (2011, February). Accelerating reservoir simulators using GPU technology. SPE Reservoir Simulation Symposium. Society of Petroleum Engineers.
Cao, H., Zaydullin, R., Liao, T., Gohaud, N., Obi, E., & Darche, G. (2021, October). Adding GPU acceleration to an industrial CPU-based simulator: Development strategy and results. SPE Reservoir Simulation Conference. Society of Petroleum Engineers.
Luo, D., Qiao, C., Gu, Y., & Zhang, J. (2020). tNavigator: A fine reservoir simulator based on modern CPU and GPU computing platforms assisting efficient development of large oil and gas fields. Computer Knowledge and Technology, 16(18), 205–206.
Mohajeri, S., Eslahi, R., Bakhtiari, M., Alizadeh, A., Zeinali, M., Madani, M., Rajabi, H., Sharifi, E., Mortezazadeh, E., & Mahdavifar, Y. (2020, November). An adaptive CPR-AMG based linear solver for simulating geometrically complicated and fractured reservoirs. Abu Dhabi International Petroleum Exhibition & Conference. Society of Petroleum Engineers.
Chen, Y., Zhang, D., Cui, S., & Li, J. (2020). Research on heterogeneous parallel algorithms for numerical simulation of dual-porosity reservoirs. Computer Engineering and Science, 42(10), 1880–1886.
Zhang, K., Chen, Z., & Liu, H. (2022). GPU-accelerated EDFM for simulation of fractured reservoirs. Journal of Petroleum Science and Engineering, 208, 109358.
Lian, P., Ji, B., Duan, T., & Zhao, H. (2020). CPU and GPU hybrid parallel numerical simulation for large complex reservoirs. China Sciencepaper, 15(5), 537–541.
Bogachev, K., Milyutin, S., Telishev, A., Nazarov, V., Shelkov, V., & Eydinov, D. (2018). High-performance reservoir simulations on modern CPU-GPU computational platforms. Rock Flow Dynamics Technical Report.
Wu, S., Li, J., Zhang, Y., & Chen, Z. (2021). Challenges in parallel reservoir simulation. SPE Journal, 26(5), 2345–2356.
Yang, X., Zhang, D., Liu, D., Cao, L., Qu, C., Liu, C., & Yu, W. (2025, March). AI-driven geology-engineering workflow for hydraulic fracture evaluation with GPU acceleration. SPE Reservoir Simulation Conference. Society of Petroleum Engineers.
Lin, M., Tayir, Zou, J., Jing, S., & Guan, Y. (2013). Application prospects of GPU computing in oil and gas exploration. Computer Systems & Applications, 22(3), 6–10.
Bernaschi, M., Celestini, A., D’Ambra, P., & Vella, F. (2023). Multi-GPU aggregation-based AMG preconditioner for iterative linear solvers. arXiv .
Rao, X., He, X., Du, K., Kwak, H., Yousef, A., & Hoteit, H. (2024). A novel projection-based embedded discrete fracture model (pEDFM) for anisotropic two-phase flow simulation using hybrid of two-point flux approximation and mimetic finite difference (TPFA-MFD) methods. Journal of Computational Physics, 499, 112736.
Rao, X., Guo, S., He, X., Kwak, H., Yousef, A., & Hoteit, H. (2025). A first streamline-based simulation method within the projection-based embedded discrete fracture model (pEDFM). Computers and Geotechnics, 185, 107357.
Rao, X., He, X., Xu, Y., Kwak, H., & Hoteit, H. (2024). Numerical simulation of carbon dioxide flooding in fractured reservoirs using generic projection-based embedded discrete fracture model. Physics of Fluids, 36(10).
Rao, X. (2023). A generic workflow of projection-based embedded discrete fracture model for flow simulation in porous media. Computational Geosciences, 27, 561–590.
Rao, X., He, X., Kwak, H., & Hoteit, H. (2024). A hybrid method combining mimetic finite difference and discontinuous Galerkin for two-phase reservoir flow problems. International Journal for Numerical Methods in Fluids.
Rao, X., Guo, S., He, X., Kwak, H., Yousef, A., & Hoteit, H. (2024). Hybrid mimetic finite difference and streamline methods for numerical simulation of two-phase flow in fractured reservoirs. Computers and Geotechnics, 166, 106048.
Rao, X., Zhao, H., & Liu, Y. (2022). A meshless numerical modeling method for fractured reservoirs based on extended finite volume method. SPE Journal, 1–40.
Rao, X., Xu, Y., & Liu, W. (n.d.). Meshless extended finite volume method for compositional reservoir models. Chinese Journal of Computational Mechanics, 1–9.
Rao, X., Zhao, H., & Liu, Y. (2023). A novel meshless method based on the virtual construction of node control domains for porous flow problems. Engineering with Computers, 1–41.
Rao, X. (2022). An upwind generalized finite difference method (GFDM) for meshless analysis of heat and mass transfer in porous media. Computational Particle Mechanics, 1–22.
