A practical implementation of acoustic full waveform inversion on graphical processing units
Abstract
Recently, Full Waveform Inversion (FWI) has gained more attention in the exploration geophysics community as a data fitting method that provides high-resolution seismic velocity models. Some of FWI essential components are a cost function to measure the misfit between observed and modeled data, a wave propagator to compute the modeled data and an initial velocity model that is iteratively updated until an acceptable decrease of the cost function is reached.
Since FWI is a wave equation based method, the computational costs are elevated. In this paper, it is presented a fast Graphical Processing Unit (GPU) FWI implementation that uses a 2D acoustic wave propagator in time and updates the model using the gradient of the cost function, which is efficiently computed with the adjoint state method. The proposed parallel implementation is tested using the Marmousi velocity model. The performance of the proposed implementation is evaluated using the NVIDIA GeForce GTX 860 GPU and compared to a serial Central Processing Unit (CPU) implementation, in terms of execution time. We also evaluate the GPU occupancy and analyze the memory requirements. Our tests show that the GPU implementation can achieve a speed-up of 26.89 times when compared to its serial CPU implementation.
References
https://doi.org/10.1190/1.1440470
Berenger, J. P. (1994). A perfectly matched layer for the absorption of electromagnetic waves. J. Comput. Phy., 114(2), 185-200.
https://doi.org/10.1006/jcph.1994.1159
Bunks, C., Saleck, F. M., Zaleski, S. & Chavent, G. (1995). Multiscale seismic waveform inversion. Geophysics, 60(5), 1457-1473.
https://doi.org/10.1190/1.1443880
Cao, D. & Liao, W. (2014). An adjoint-based hybrid compu- tational method for crosswell seismic inversion. Comput. Sci. Eng., 16(6), 60-67.
https://doi.org/10.1109/MCSE.2014.1
Collino, F. & Tsogka, C. (2001). Application of the perfectly matched absorbing layer model to the linear elastodynamic problem in anisotropic heterogeneous media. Geophysics, 66(1), 294-307.
https://doi.org/10.1190/1.1444908
Etienne, V., Tonellot, T., Thierry, P., Berthoumieux, V. & Andreolli, C. (2014). Speeding-up FWI by one order of magnitude. EAGE Workshop on High Performance Computing for Upstream. Chania, Crete.
https://doi.org/10.3997/2214-4609.20141905
Hu, W., Abubakar, A. & Habashy, T. M. (2007). Application of the nearly perfectly matched layer in acoustic wave modeling. Geophysics, 72(5), 169-175.
https://doi.org/10.1190/1.2738553
Kim, Y., Shin, C. & Calandra, H. (2012). 3D hybrid waveform inversion with GPU devices. SEG Annual Meeting. Las Vegas, Nevada. SEG-2012-0138.
https://doi.org/10.1190/segam2012-0138.1
Lailly, P. (1983). The seismic inverse problem as a sequence of before stack migrations. Conference on Inverse Scattering: Theory and Application. Society for Industrial and Applied Mathematics, Philadelphia, PA.
Mao, J., Wu, R.S. & Wang, B. (2012). Multiscale full waveform inversion using GPU. SEG Annual Meeting. Las Vegas, Nevada, SEG-2012-2575.
https://doi.org/10.1190/segam2012-0575.1
Pasalic, D. & McGarry, R. (2010). Convolutional perfectly matched layer for isotropic and anisotropic acoustic wave equations. SEG Annual Meeting. Denver, Colorado, SEG- 2010-2925.
https://doi.org/10.1190/1.3513453
Plessix, R. E. (2006). A review of the adjoint-state method for computing the gradient of a functional with geophysical applications. Geophys. J. Int., 167(2), 495-503.
https://doi.org/10.1111/j.1365-246X.2006.02978.x
Tarantola, A. (1984). Inversion of seismic reflection data in the acoustic approximation. Geophysics, 49(8), 1259-1266.
https://doi.org/10.1190/1.1441754
Tarantola, A. (1987). Inverse problem theory: Methods for data fitting and model parameter estimation. Elsevier.
Thierry, P., Donno, D. & Noble, M. (2014). Fast 2D FWI on a multi and many-cores workstation. EGU General Assembly, 16: Vienna, Austria.
Versteeg, R. & Grau, G. (1991). The Marmousi experience: Proceedings of the 1990 EAEG Workshop on practical aspects of seismic data inversion.
Vigh, D., Cheng, X., Jiao, K., Sun, D. & Kapoor, J. (2014). Multiparameter TTI full waveform inversion on long- offset broadband acquisition: A case of study. SEG Annual Meeting. Denver, Colorado, SEG-2014-0530.
https://doi.org/10.1190/segam2014-0530.1
Vir eux, J. & Operto, S. (2009). An overview of full-waveform inversion in exploration geophysics. Geophysics, 74(6), 1-26.
https://doi.org/10.1190/1.3238367
Wang, B., Gao, J., Zhang, H. & Zhao,W. (2011). CUDA-based acceleration of full waveform inversion on GPU. SEG Annual Meeting. San Antonio, Texas, SEG-2011-2528.
https://doi.org/10.1190/1.3627717
Weiss, R. M. & Shragge, J. (2013). Solving 3D anisotropic elastic wave equations on parallel GPU devices. Geophysics, 78(2), 7-15.
https://doi.org/10.1190/geo2012-0063.1
Yang, P., Gao, J. & Wang, B. (2015). A graphics processing unit implementation of time-domain full-waveform inversion. Geophysics, 80(3), 31-39.
https://doi.org/10.1190/geo2014-0283.1
Zhang, M., Sui, Z., Wang, H., Ren, H., Wang, Y. & Meng, X. (2014). An approach of full waveform inversion on GPU clusters and its application on land datasets. 76th EAGE Conference and Exhibition. Amsterdam.
https://doi.org/10.3997/2214-4609.20141292
Downloads
