Advanced Imaging


Advanced Imaging Magazine

Updated: January 12th, 2011 10:01 AM CDT

Going to the Well

Accelerated number crunching makes oil and gas exploration more efficient
Integrated computation and visualization workflow
© Mercury Computer Systems
Integrated computation and visualization workflow using NVIDIA Tesla and Open Inventor from Mercury Computer Systems. Mercury Computer Systems
Input data
© Mercury Computer Systems
Input data: Seismic amplitude volume managed and visualized using Open Inventor.
Output data
© Mercury Computer Systems
Output data: Seismic attribute volume managed and visualized using Open Inventor.
NVIDIA's Tesla GPU server.
NVIDIA's Tesla GPU server.

By Barry Hochfelder

Some jobs are just too big for one person, one company or one technology. In oil and gas exploration, for example, one seismic survey can equal 10 terabytes of data. (That's one trillion bytes or 1,000 megabytes.) Information can require hundreds of teraflops (floating point operations per second) of computer capability and petabytes (a single petabyte is one quadrillion bytes) of storage.

To help deal with all that data, Mercury Computer Systems, Inc. (Chelmsford, Mass.) and NVIDIA Corp. (Santa Clara, Calif.) are collaborating to provide exploration and production (E&P) application developers with a comprehensive solution for simultaneous computation and visualization of huge 3D seismic data sets—or any highly demanding computing tasks in the interpretation and simulation workflows.

Mercury will integrate NVIDIA's Tesla™ GPU computing power inside its Open Inventor™ 3D development toolkit for highly demanding computing tasks in the data analysis, interpretation, and simulation E&P workflows. E&P application solutions will get interoperability between 3D visualization and the computation on the fly. Open Inventor has been largely adopted by the Oil E&P industry. It includes VolumeViz LDM, a high-performance volume visualization technology for extremely large seismic data sets (pre- and post-stacks), and ReservoirViz LDM, advanced visualization components for reservoir modeling and simulation.

Tesla is a high-performance GPU computing solution that delivers supercomputing power to any CPU-based workstation or server. The CUDA™ C-Compiler allows optimized programming of the GPU in the C language, turning the parallel processing capabilities of the GPU into massive computing power. Open Inventor™ is an object-oriented, cross-platform 3D graphics toolkit for rapid prototyping and development of industrial-strength interactive 3D applications using C++, .NET or Java. It provides the power and functionality of OpenGL at an object-oriented level, including advanced support of OpenGL shaders to further enhance the 3D visualization perception. It allows users to implement rendering distribution on clusters, scene and image compositing to visualize huge amounts of data remotely, and a broad range of visualization techniques on 2D and 3D meshes for reservoir modeling and simulation.

During the most recent Supercomputing show in Reno, Nev., Mercury showed a prototype application developed with Ocean, an application development framework for extending the Schlumberger Petrel seismic-to-simulation software. Schlumberger Petrel demonstrates seamless integration between high-performance computing and high-quality visualization for very large 3D seismic data sets. Both Petrel and Ocean are built on top of the Open Inventor and VolumeViz™ LDM technologies. This new computing offering allows software developers to integrate high-performance computing CUDA code through the Open Inventor toolkit.

1 2 3 4 next

Subscribe to our RSS Feeds