Tempest MORE is a high-resolution and high-performance full field black oil, compositional and thermal reservoir simulator. Designed for models with millions of cells and thousands of wells, it's essential that the programme delivers swift and reliable throughput.
"In the reservoir simulation market, the most critical feature operators are looking for is speed", explained Michele Taroni, Principal Software Developer of Roxar Software Solutions. "As models get progressively larger they need to be run faster and in parallel, which make speed and scalability absolute musts."
In order to meet this demand, Emerson invested in Allinea Forge, a complete toolset designed for developers of compute intensive and parallel code. Within that toolset is Allinea DDT, a powerful and intuitive debugger that offers unrivalled support for debugging multi-process and multi-threaded applications. Complex interacting components can therefore be debugged quickly and easily, saving precious development time.
Allinea Forge also contains Allinea's profiler for multi-process and multi-threaded C, C++ and F90 codes - Allinea MAP. Providing in-depth analysis and bottleneck pinpointing to the source line, Allinea MAP enables developers to drill deep down into the performance of their code.
"Allinea DDT and Allinea MAP both had an immediate impact on the performance of our code", stated Roxar Software Solutions developer Michele Taroni.
"In one user case alone we were not only able to discover what was behind a slowdown, but we were then able to identify and prevent another bottleneck from happening. As a result, the model is now running around 30 per cent faster than it was before."
David Lecomber, CEO of Allinea Software, added: "It's fantastic to see Allinea Forge having a rapid impact on the software challenges of the oil and gas market. Understanding and maximizing the performance of reservoirs through the lifecycle technologies developed by Emerson's Roxar Software Solutions benefits everyone. The ability of our tools to profile and debug these complex parallel codes is accelerating the performance of code and speeding up the development itself."