Livermore Computing

Providing high-performance computing resources to Defense Systems
and other organizations at the Lawrence Livermore National Laboratory

The Livermore Computing organization has been a leader in the development of High Performance computing since the Lawrence Livermore National Laboratory (LLNL) was founded in 1952. During the Cold War, Livermore Computing concentrated on supporting the development of our nation's nuclear deterrent, designing early supercomputers, like the LARC (Livermore Advanced Research Computer), and developing the first time-sharing system in the 1960s and 1970s, large archival storage systems in the 1970s and 1980s, and distributed systems in the 1980s and 1990s.

The dramatic changes taking place in the world's political climate are driving major changes in LLNL's missions. As a result, Livermore Computing is devoting increasing amounts of effort to such tasks as ensuring a safe and reliable nuclear stockpile, supporting non-proliferation and counter-proliferation activities, and developing and transferring to industry advanced technologies that improve U.S. economic competitiveness.

Today, Livermore Computing maintains classified and unclassified networks to support Defense Systems and other LLNL research organizations. These facilities, which have traditionally included conventional and vector supercomputers, are being enhanced with workstation clusters and massively parallel machines incorporating the latest, most cost-effective technology. The newest machine is a Meiko CS-2 parallel processor. This machine, when fully configured, will contain 256 computational nodes, each of which is equivalent to the supercomputers of just a few years ago, with a total memory capacity of 32 billion bytes. The Livermore Computing computational power is supported by a standards-based software environment, FDDI and HIPPI networks, and a disk and tape archive giving on-line access to fifteen trillion bytes of data.

Collaborations

• IEEE Metadata Workshops -- Providing an information source for metadata and data management issues.

• National Storage Laboratory -- Investigating, demonstrating, and commercializing high-performance storage environments.

• Parallel Tools Consortium -- Defining, developing, and promoting parallel, portable, scalable software tools.

Current Projects

• Facility for Advanced Scalable computing Technology (FAST) -- Providing an unclassified world-class, high-performance computing environment.

• Full Spectrum Computing -- Providing a wide range of computing resources to LLNL scientists.

• Gas and Oil NII Testbed -- Applying the computational research and system integration abilities to U.S. industry issues.

• Intelligent Archive -- Providing intelligent searching, organizing, and browsing capabilities for LLNL scientists' data and information.

• Mathematical Software Service -- Solving mathematical problems more accurately at lower cost.

• Massively Parallel Processing -- Defining and implementing a large-scale, parallel computing environment.

• Meiko Integration -- Incorporating the Meiko CS-2 into the Livermore Computing environment.

• Meiko Parallel Development Environment -- Working to provide robust, scalable application development environment tools for the Meiko CS-2.

• Mesh-TV -- Developing an interactive graphical analysis tool for visualizing and analyzing 2D and 3D mesh data.

• Production Control System -- Developing resource allocation mechanisms for Unix environments.

• Scalable I/O Facility -- Building the next generation parallel I/O infrastructure for high performance applications.

• Scientific Visualization -- Providing visualization expertise through the Visualization Laboratory and a variety of research and development efforts.

Available Software

• ELROS -- An Embedded Language for Remote Operations Service

• LLNL XDIR -- A graphical user interface for file transfer and for direct manipulation of local and remote directories on UNIX and other platforms.

• LLNL XFTP -- An X Window FTP Client.

If you have questions about the Livermore Computing organization, contact:

Go to the LLNL Home Page
LLNL Disclaimers

UCRL: TB-115079-2