Skip to main content
NASA Advanced Supercomputing (NAS) Division World-class modeling, simulation, and supercomputing in support of the NASA mission

Application Porting and Optimization

NAS applications experts work closely with researchers supporting NASA missions to ensure that their applications run smoothly on the high-performance computing systems—often working with groups who use the most computing resources to verify the accuracy of their code performance. Our experts help users with all aspects of their projects, including troubleshooting issues, porting codes to NAS systems, developing customized scripts to parallelize jobs and applications, conducting extensive code analysis and re-engineering to enhance performance, and providing training. In addition, the team develops custom tools that help users monitor and schedule jobs, analyze codes, and optimize performance.

Image of radiation and plasma environment for satellites orbiting Earth
Working with the Application Optimization team, heliophysics researchers at Auburn University were able to more than double their code performance, enhancing their capability to predict and accurately specify the radiation and plasma environment for satellites orbiting the Earth.

Proactive monitoring of system components can also lead to substantial improvements for users, helping them to meet their NASA mission milestones. In one recent case, filesystem monitoring helped identify inefficient I/O patterns in a code used to analyze the Avcoat heat shield system that will protect NASA’s Orion spacecraft during Earth re-entry. The application optimization team developed a more efficient approach that reduced overall time-to-solution by a factor of five—enabling the researchers to complete their analysis in time for Orion’s Probability Risk Assessment for the Artemis I mission.

In addition to supporting NAS users, the application optimization team is responsible for benchmarking new hardware when new supercomputers are deployed or when new processor types are added to existing systems, in order to determine the Standard Billing Unit (SBU) rates that are used to track usage. The team also ensures that the SBU benchmark suite is representative of typical workloads.