Oil and Gas

Seismic Processing and Reservoir Modeling

The upstream oil and gas industry relies on high performance computing software.

Seismic data processing in exploration requires vast quantities of compute time. Seismic shots recorded from ocean-based survey vessels or land-based surveys take hours on hundreds of processors to generate a single image of the subsurface.

A typical survey creates thousands of shots to analyze – and it is important to be able to apply accurate methods fast. Today's methods such as full wave inversion (FWI) or reverse time migration (RTM) are far more intensive than previous methods.

External seismic data-processing houses and in-house processing and quality control (QC) teams use software developed by computational geoscientists. Faster and more scalable processing reduces the time to analyze interesting geologies and improves well-location decisions – potentially saving millions of dollars.

Reservoir modeling enables the extraction of greater proportions of oil and gas from existing reservoirs. Computational fluid dynamics (CFD) methods model the movement of fluids and enable the modeling of the lifecycle of a reservoir. This modeling ensures that extraction is orchestrated for optimal yield. The models perform long running calculations on thousands of processors. Fast throughput is critical for the industry.

Using Allinea Forge enables developers to accelerate their software by using more processors concurrently and using existing processors more efficiently - and Allinea Performance Reports can ensure that an application remains efficient once it is deployed.

Our Success Stories

Tullow Oil

Tullow Oil is a leading independent oil and gas exploration and production company with a focus on finding and monetizing oil in Africa and the Atlantic margins.

Software developed in-house to analyze seismic survey data is used to accurately identify oil and gas reserves in the subsurface. Increasing the performance of that software directly translates to faster turnaround of important results, and ultimately, more confidence in well location decisions.

Working with the Irish Centre for High End Computing (ICHEC), Tullow Oil used Allinea Forge and Allinea Performance Reports to provide valuable insight in the performance optimization process, reducing development time and costs.

Allinea’s tools helped identify the bugs and bottlenecks that arose in the development and optimization of TORTIA. The product of that development delivers high quality images in a fraction of the time.
Dr Sean Delaney, Computational Geophysicist, Tullow Oil

CGG - commercial geosciences

CGG, one of the world's leading commercial geosciences companies says Allinea DDT helps them to solve bugs up to five times faster than before.

With clients relying on their high-quality and on-schedule releases, it's the globally distributed software team’s mission – and resolving bugs quickly is critical to success.  Allinea DDT, the debugger for multi-threaded and MPI parallel code provides them with the ability to deliver.

The first time I used it was to debug an MPI code and I found it was the only tool that could solve the problem.
Daniel Trad, an experienced researcher, CGG

Allinea Forge - the C, C++, F90 development tool suite
Emerson Roxar

Speed and quality are critical to reservoir simulation software and to the development teams behind it. Allinea Forge had immediate impact on Roxar Software Solutions in the race for higher resolution.

Clients need higher precision results faster - forcing a huge increase in the number of compute cores used per simulation - but high core counts bring more complex performance bottlenecks and bugs that conventional methods cannot solve quickly.

With Allinea Forge, the development team at Emerson Roxar are enabling faster simulations for clients, faster resolution of problems - and getting more time to develop software.

Allinea Forge had an immediate impact on the performance of our code. In one case, we not only removed a slowdown, but we were able to identify and prevent a bottleneck from happening. The model is now running 30% faster than it was before.
Michele Taroni, Principal Software Developer