Technologies for Large-Scale Numerical Simulation
- Created on Tuesday, 01 November 2011
NASA Ames Research Center is home to the world’s seventh most powerful supercomputer, Pleiades. Named after the star cluster, it represents NASA’s state-of-the-art technology for meeting the agency’s supercomputing requirements, enabling NASA scientist s and engineers to conduct modeling and simulation for NASA missions. As one of the most powerful general-purpose supercomputers ever built, Pleiades has the computing power equivalent to that of 41,000 MacBook Pros.
At Ames, where complex mathematical
models and huge data sets are the norm,
Pleiades goes through constant expansions
and upgrades to help solve problems
across NASA’s mission areas: aeronautics
research, human exploration and
operations, and Earth and space science.
NASA scientists and engineers are
increasingly turning to large-scale numerical
simulation on supercomputers to
advance their understanding of complex
aerospace, Earth, and astrophysical systems.
To meet these demands, efforts are
underway to improve the productivity of
NASA’s supercomputing users, as well as
to accelerate advancement of NASA science
and engineering at the Center.
NASA’s high-end computing (HEC) facilities are continually challenged to purchase, power, and cool sufficient HEC resources to keep up with equally rapidly growing user demands. As a result, the Center is constantly looking for more efficient and effective supercomputing approaches to supply increased HEC capability or capacity per dollar and/or per Watt for NASA applications. Examples include:
• Novel computational accelerators and architectures.
• Enhanced visualization technologies.
• Improved algorithms for heavily used application software.
• Power-aware “green” computing technologies and techniques.
• Systems (including both hardware and software) for data-intensive computing.
• Approaches to effectively manage and utilize many-core processors, including algorithmic changes, compiler techniques, and runtime systems.
Improving the user’s productivity environment is also critical. What is needed are more intuitive, intelligent, user-customizable, and integrated interfaces to supercomputing resources, enabling users to more completely leverage the power of HEC to increase their productivity. Such interfaces could enhance many essential supercomputing tasks, including:
• Accessing and managing resources
• Getting services
• Developing software (e.g., debugging and performance analysis)
• Running computations
• Managing files and data
• Analyzing and visualizing results
• Transmitting data
HEC interfaces and environments designed to run on handheld devices and tablets are also of interest.
Cloud computing has demonstrated tremendous success in providing easy access to computing resources, data storage, and widely used software. Officials at Ames are seeking technologies that would enable beneficial features of cloud computing to be offered as part of NASA’s supercomputing operations and services. Such beneficial features might include more open account granting, user-customized software environments, support for long-running jobs (e.g., Web applications and services) , real-time access to external Web resources and services, and reduced job failure due to system issues.