The neutron transport code UNIC being developed by a team of computer scientists and nuclear engineers at Argonne National Laboratory enables researchers to acquire a fine-grained model of a nuclear reactor core for the first time. “The UNIC code is intended to reduce the uncertainties and biases in reactor design calculations by progressively replacing existing multilevel averaging techniques with more direct solution methods based on explicit reactor geometries,” says Argonne scientist Andrew Siegel. The Argonne researchers have executed detailed simulations of the Zero Power Reactor experiments on as many as 163,840 processor cores of the Blue Gene/P and 222,912 processor cores of the Cray XT5 at Oak Ridge National Laboratory, as well as on 294,912 processor cores of a Blue Gene/P at Germany’s Julich Supercomputing Center. UNIC has enabled researchers to successfully represent the details of the full reactor geometry as well as compare the results directly with the experimental data. The scientists say the code could play an essential role in the development of safe, affordable, and green nuclear reactors. UNIC gives researchers a better understanding of the behavior of existing reactor systems and also allows them to anticipate the behavior of many newly proposed systems with untested design characteristics.

View Full Article

For More Information Visit: http://www.cpccci.com

European researchers have devised a new software development paradigm using an assembly line-style development process. “Think of this as a sandwich shop, where you have different products coming from a product line that shares ingredients, which customers can pick and choose,” says AMPLE project coordinator and Lancaster University professor Awais Rashid. The asset base features modular software elements that establish a Software Product Line (SPL), within which is managed the entire software lifecycle from design and development through deployment and maintenance. The AMPLE team created analyses tools that guide users on system development. Rashid says the results of the AMPLE tool analyses match those of human software experts, but the AMPLE software is capable of much faster assessment and can be used by non-experts. Other tools in the chain let companies generate their own modular software components, to put them together for a specific job, and to test and validate the resulting application. Another key element is the maintenance, repair, and modification of both the SPL and the software it creates.

View Full Article

For More Information Visit: http://www.cpccci.com

The computer industry had an opportunity to learn about the technical details of Google’s infrastructure during LADIS 2009, ACM’s recent SIGOPS International Workshop on Large Scale Distributed Systems and Middleware. Jeff Dean, a Google engineer who was one of the keynote speakers, also talked about Spanner, a new storage and computation system for automating the management of services across multiple data centers. Spanner, which will have a scale of 1 million to 10 million servers in the future, would be capable of automatically allocating resources across “entire fleets of machines,” Dean says. The goal will be “automatic, dynamic, worldwide placement of data and computation to minimize latency or cost.” Spanner also would offer a cost management strategy for addressing regional differences in bandwidth and power costs. Google would have energy management opportunities because Spanner can seamlessly shift workloads between data centers. Automated capacity management also would enable Google to route around failures or data center downtime as well as plan more energy-efficient facilities.

View Full Article

For More Information Visit: http://www.cpccci.com