Altair > Case Studies > Optimizing Engineering Cluster Performance at GE's Oil & Gas Business

Optimizing Engineering Cluster Performance at GE's Oil & Gas Business

Technology Category

Sensors - Gas Sensors
Sensors - Utility Meters

Applicable Industries

Oil & Gas
Transportation

Applicable Functions

Logistics & Transportation

Use Cases

Inventory Management
Time Sensitive Networking

Services

System Integration

About The Customer

Nuovo Pignone, now a key technological component of GE's Oil & Gas business, began life 100 years ago as a foundry. Today, more than 20,000 machines manufactured by this GE business are operated worldwide by major companies to keep petroleum products moving from oil rigs to storage facilities to refineries to distribution points. The engineering groups at GE's Oil & Gas Florence headquarters work extensively with commercial and proprietary CFD and CEA tools which are used to build and test product designs. The gas turbine engineering group designs gas turbines that are installed in the pipeline system or offshore for mechanical drive applications. Their customers are major petroleum companies like Shell and Texaco, with sites around the world.

The Challenge

GE's Oil & Gas business was facing a significant challenge in job scheduling for their applications. The gas turbine engineering group was upgrading its High-Performance Computing (HPC) resources, moving from an 8-CPU UNIX server to a 20-CPU HP ProLiant 380 cluster running Red Hat Linux. However, they encountered unexpected difficulties in job scheduling using open source code. The situation was becoming increasingly difficult to manage, causing delays in production. The group leader, Alessandro Ciani, was tasked with getting the cluster into production but was frustrated with the delays. The situation was urgent, and a solution was needed to ensure the system could be put into production.

The Solution

Altair, a global technology company providing solutions in product development, high-performance computing and data intelligence, was approached to provide a solution. Instead of a standard presentation, Altair suggested a pilot installation of their PBS Professional® workload management software. This integrated workload management solution was optimized for running the StarCD application. The GE engineering team was soon able to submit and manage its StarCD workload with confidence. Following the success of the pilot, GE's Oil & Gas business purchased PBS Professional licenses for all CPUs in the cluster and hired Altair to integrate their other applications into their PBS Professional cluster. The cluster was expanded by adding heterogeneous nodes, and all 90 processors in the cluster and the corresponding application licenses were efficiently managed with PBS Professional.

Operational Impact

The implementation of Altair's PBS Professional® workload management software has significantly improved the efficiency and productivity of GE's Oil & Gas business. The engineering team can now submit and manage its StarCD workload with confidence, reducing delays and increasing productivity. The cluster has been expanded and is now able to handle a larger workload. The solution has also provided the opportunity for future improvements, such as harvesting unused compute cycles from engineering workstations and resource-sharing with another of GE's Oil & Gas engineering groups. The success of this solution has led to considerations for PBS Professional to become an enterprise-wide solution, further enhancing the company's operational efficiency.

Quantitative Benefit

Expanded the cluster by adding heterogeneous nodes, increasing the total to 45 two-processor nodes.
Efficiently managed all 90 processors in the cluster and the corresponding application licenses with PBS Professional.
15 to 20 engineers are able to run jobs on the gas turbine engineering cluster from their desktop PCs simultaneously.