October 2013 Summaries

Rescale participated in the Society of Petroleum Engineers' Annual Technical Conference and Exhibition, the largest event in oil and gas exploration and production, held in New Orleans. The conference highlighted the use of high-performance computing (HPC) in various aspects of exploration and production (E&P), including seismic processing, reservoir modeling, and meteorology. These applications often require significant computing power, which is increasingly moving to cloud-based solutions due to the flexibility and cost-effectiveness they offer compared to traditional in-house clusters. This transition allows for more customizable and efficient resource use, as well as innovative pay-per-use licensing models. Rescale, which provides a pay-per-use HPC platform, is actively engaging with simulation software vendors in the industry to support this shift, positioning itself as a leader in cloud-based computing for E&P processes.

In the oil and gas industry, complex and resource-intensive simulations are crucial for tasks like reservoir modeling and drilling applications, often requiring high-performance computing (HPC) due to their demanding nature. A leading oilfield service company addressed these challenges by leveraging Rescale's secure, web-based platform to conduct simulations using LS-DYNA software, allowing them to efficiently evaluate acoustic emission properties across different drilling scenarios. By running simulations on 16, 32, and 64 cores, Rescale's dynamic provisioning reduced computation times significantly compared to the company's in-house cluster, achieving a 13% reduction in compute time on 16 cores and more than 75% faster results with 64 cores, while also providing cost savings of over $6,000. This not only shortened convergence times but also uncovered new insights into drilling scenarios, aiding in the determination of optimal acoustic emissions environments.

At the ISC Cloud’13 Conference in Heidelberg, Germany, discussions centered around the growing integration of high-performance computing (HPC) in cloud environments, with particular emphasis on engineering and scientific simulations. Attendees included representatives from major manufacturers, software vendors, and end users across various sectors, all exploring innovative cloud solutions like virtual real-time markets for HPC capacity and hybrid cloud setups for critical financial services. The conference highlighted a strong interest in cloud-based engineering simulations, especially among smaller companies seeking cost-effective solutions, while large enterprises showed interest in the cloud's burst capabilities. Key takeaways from the event included the importance of business models and data security in cloud adoption, as well as the recognition by engineering software vendors that cloud HPC represents a significant opportunity for growth and innovation. The event provided valuable insights into the market's current state and future potential, signaling a promising outlook for cloud-based simulations.

Stanford University Unstructured (SU2) is an open-source C++ toolset designed for solving Partial Differential Equations (PDEs), particularly useful in PDE-constrained optimization problems across various fields, including aerodynamics and electrodynamics. Developed by Stanford's Aerospace Design Laboratory, SU2 is valuable in academia and industry for computational analysis, a sector often limited by expensive proprietary codes and resource constraints. SU2 is offered alongside Rescale's scalable computing platform to provide a cost-effective, accessible solution. A demonstration of SU2's capabilities involved an aerodynamic shape optimization of the ONERA M6 wing model, achieving a 24.6% reduction in pressure drag through iterative design cycles. This process involved solving direct flow and continuous adjoint problems to compute gradients for optimal shape adjustments, while maintaining the lift coefficient to manage induced drag. The results, including a significant reduction in pressure drag, highlight SU2's effectiveness in practical aerodynamic applications.