1887

Abstract

Summary

Uses of artificial-intelligence/machine-learning (AI/ML) applications for data processing and decision making in internet of things (IoT) devices have proliferated in recent years, spurring development of new hardware suitable for inference in edge environments. The Nvidia Jetson Orin system-on-chip is one such device, featuring a 1024-core Ampere architecture GPU, whilst drawing a mere 25W. This hardware highlights the possibility of running wave-equation-based seismic workflows such as RTM and FWI closer to the edge.

This work benchmarks the throughput of the Jetson Orin for a range of wave equation kernels commonly used in seismic processing. Throughput per unit power consumption is also explored, and the aforementioned metrics are compared with those for a reference workstation/server CPU architecture. The Nvidia Jetson Orin achieves respectable performance across a range of wavesolver kernels used in workflows such as RTM and FWI, with much greater performance-per-Watt than conventional architectures.

Furthermore, the implications of such performance metrics are considered in the wider context of seismic imaging trends, particularly the introduction of disruptive technologies to streamline acquisition and processing, and development of novel imaging strategies. It is anticipated that, as datasets grow and timelines shorten, effective leveraging of edge compute will provide substantial advantage for future seismic imaging.

© EAGE Publications BV
Loading

Article metrics loading...

/content/papers/10.3997/2214-4609.2025643015
2025-10-06
2026-02-15

  • From This Site

    /content/papers/10.3997/2214-4609.2025643015
    dcterms_title,dcterms_subject,pub_keyword
    -contentType:Journal -contentType:Contributor -contentType:Concept -contentType:Institution
    10
    5
Loading full text...

Full text loading...

References

  1. Cui, Laizhong et al., “A survey on application of machine learning for Internet of Things.” International Journal of Machine Learning and Cybernetics9 (2018): 1399–1417.
     [Google Scholar]
  2. Samie, Farzad, LarsBauer, and JörgHenkel. “From cloud down to things: An overview of machine learning in internet of things.” IEEE Internet of Things Journal6.3 (2019): 4921–4934.
     [Google Scholar]
  3. Fabien-Ouellet, Gabriel. “Seismic modeling and inversion using half-precision floating-point numbers.” Geophysics85.3 (2020): F65–F76.
     [Google Scholar]
  4. Liu, Xiaoyan et al., “Toward accelerated stencil computation by adapting tensor core unit on gpu.” Proceedings of the 36th ACM International Conference on Supercomputing. 2022.
     [Google Scholar]
  5. “NVIDIA Jetson AGX Orin.” NVIDIA, www.nvidia.com/en-gb/autonomous-machines/embedded-systems/jetson-orin.
     [Google Scholar]
  6. Louboutin, Mathias et al., “Devito (v3. 1.0): an embedded domain-specific language for finite differences and geophysical exploration.” Geoscientific Model Development12.3 (2019): 1165–1187.
     [Google Scholar]
  7. Luporini, Fabio et al., “Architecture and performance of Devito, a system for automated stencil computation.” ACM Transactions on Mathematical Software (TOMS)46.1 (2020): 1–28.
     [Google Scholar]
  8. Lan, Bo et al., “Seismic Phase Picking Using a Cross-Attention Network on NVIDIA Jetson Xavier NX.” IEEE Access (2024).
     [Google Scholar]
  9. Shin, Ban-Sok, LuisWientgens, and DmitriyShutin. “Imaging on the Edge: GPU-Accelerated Traveltime Tomography on a Jetson Nano.” 2024 32nd European Signal Processing Conference (EUSIPCO). IEEE, 2024.
     [Google Scholar]
  10. Sarkar, Subhadeep, SubarnaChatterjee, and SudipMisra. “Assessment of the suitability of fog computing in the context of internet of things.” IEEE Transactions on Cloud Computing6.1 (2015): 46–59.
     [Google Scholar]
  11. Ourabah, Amine, and AllanChatenay. “Unlocking ultra-high-density seismic for CCUS applications by combining nimble nodes and agile source technologies.” The Leading Edge41.1 (2022): 27–33.
     [Google Scholar]
  12. Shin, Ban-Sok, et al. “Multi-agent 3D seismic exploration using adapt-then-combine full waveform inversion in a hardware-in-the-loop system.” ICASSP 2024–2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2024.
     [Google Scholar]
/content/papers/10.3997/2214-4609.2025643015
Loading
Emerging AI Hardware Enables Low-Power, High-Throughput Seismic Imaging on the Edge
Conference Proceedings 2025, 1 (2025); https://doi.org/10.3997/2214-4609.2025643015
/content/papers/10.3997/2214-4609.2025643015
/content/papers/10.3997/2214-4609.2025643015
Loading

Data & Media loading...

Most Cited This Month Most Cited RSS feed

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error