This article discusses the emergence of general computation on the graphics processing unit (GPGPU) as a powerful tool for enabling true interactive 3D seismic interpretation such that geologic features are computed, mapped, and visualized in real time. We review previous work in GPU-based methods for both prestack data and migrated volumes, but we will focus on techniques that facilitate interactive and real-time interpretation of migrated seismic volumes. The field of GPGPU technology is changing rapidly with updates frequently occurring at six-month cycles, and often bring major changes to processor architecture and programmability. In this article we provide readers with an overview of the current state of GPGPU, work being done in our industry, and discuss the greater implications these changes may have in seismic volume interpretation. We will address the unique computational challenges as well as the significant gains that can be achieved by computing on low-cost commodity GPUs.

You do not currently have access to this article.