Foster Findlay Associates


In the search for oil and gas, the geological information provided by seismic images of the earth is vital.


By interpreting the data produced by seismic imaging surveys, geoscientists can identify the likely presence of hydrocarbon reserves and understand how to extract resources most effectively.

Today, sophisticated visualisation systems and computer-aided tools are used to streamline what was previously a subjective and labour intensive process. However, the challenges facing the oil and gas industry mean that geoscientists must process increasing amounts of data as dwindling reserves require them to pinpoint smaller, more complex reservoirs with greater speed and accuracy.


UK-based company ffA provides 3D seismic analysis software and services to the oil and gas industry. Its software tools extract detailed information from 3D seismic data, providing a greater understanding of complex 3D geology and improving productivity within the interpretation process. The tools are sophisticated but also compute intensive so results can take hours or even days to produce on typical workstations systems used by the industry.

In June 2008 ffA announced a strategic partnership with NVIDIA to integrate CUDA-based graphics processing unit (GPU) computation technology into ffA’s 3D Seismic Image Processing and Analysis IPA software. The parallel nature of the task makes it extremely well suited to the GPU and ffA has been investigating GPU computation since 2007. The adoption of CUDA has enabled ffA to accelerate its R&D programme significantly. Moving to the GPU is providing up to two orders of magnitude speed increase within ffA software. This shortens computation tasks from days to minutes, dramatically increasing the amount of data that geoscientists can analyse in a given timeframe.

Steve Purves, ffA’s Technical director commented: “We have been using the graphical API successfully for some time to perform complex analysis routines with near-real time performance. CUDA now allows us to develop for the GPU more quickly, more cost effectively and with more significant improvements in performance. Above all, it promises to provide a scalable and portable way to accelerate our software applications across a wide range of PC desktop and mobile workstation configurations.”


CUDA is allowing ffA to provide scalable high performance computation for seismic data on hardware platforms equipped with NVIDIA Quadro FX GPUs that are already ubiquitous within ffA’s user base. In addition, its latest results using NVIDIA Tesla GPUs have produced performance improvements of between 40x and 98x versus the same single threaded code.

“CUDA is allowing us to accelerate our algorithms much more effectively than we would be able to do through optimisation of code for modern quad core CPUs. We intend to release a GPU enabled version of our SVI Pro and SEA 3D Pro product suites in 2009,” adds Steve Purves.

“Understanding of uncertainty and risk is critical at all stages of hydrocarbon exploration and production. By dramatically changing the way seismic interpretation is carried out, GPU computation will give geoscientists unprecedented levels of interactivity with very large volumes of data. This will lead to major improvements in productivity and impact directly on exploration success and hydrocarbon recovery.”

For more details of ffA’s results with CUDA, please visit //