One million trillion 'flops' per second targeted by new Institute for Advanced Architectures

Feb. 21, 2008

'Exascale' computing envisioned by Sandia and Oak Ridge researchers

Preparing groundwork for an exascale computer is the mission of the new Institute for Advanced Architectures, launched jointly at Sandia and Oak Ridge national laboratories.

An exaflop is a thousand times faster than a petaflop, itself a thousand times faster than a teraflop. Teraflop computers -- the first was developed 10 years ago at Sandia -- currently are the state of the art. They do trillions of calculations a second. Exaflop computers would perform a million trillion calculations per second.

The idea behind the institute -- under consideration for a year and a half prior to its opening -- is "to close critical gaps between theoretical peak performance and actual performance on current supercomputers," says Sandia project lead Sudip Dosanjh. "We believe this can be done by developing novel and innovative computer architectures."

Ultrafast supercomputers improve detection of real-world conditions by helping researchers more closely examine the interactions of larger numbers of particles over time periods divided into smaller segments.

"An exascale computer is essential to perform more accurate simulations that, in turn, support solutions for emerging science and engineering challenges in national defense, energy assurance, advanced materials, climate, and medicine," says James Peery, director of computation, computers and math.

The institute is funded in FY08 by congressional mandate at $7.4 million. It is supported by the National Nuclear Security Administration and the Department of Energy's Office of Science. Sandia is an NNSA laboratory.

One aim, Dosanjh says, is to reduce or eliminate the growing mismatch between data movement and processing speeds.

Processing speed refers to the rapidity with which a processor can manipulate data to solve its part of a larger problem. Data movement refers to the act of getting data from a computer's memory to its processing chip and then back again. The larger the machine, the farther away from a processor the data may be stored and the slower the movement of data.

"In an exascale computer, data might be tens of thousands of processors away from the processor that wants it," says Sandia computer architect Doug Doerfler. "But until that processor gets its data, it has nothing useful to do. One key to scalability is to make sure all processors have something to work on at all times."

Compounding the problem is new technology that has enabled designers to split a processor into first two, then four, and now eight cores on a single die. Some special-purpose processors have 24 or more cores on a die. Dosanjh suggests there might eventually be hundreds operating in parallel on a single chip.

"In order to continue to make progress in running scientific applications at these [very large] scales," says Jeff Nichols, who heads the Oak Ridge branch of the institute, "we need to address our ability to maintain the balance between the hardware and the software. There are huge software and programming challenges and our goal is to do the critical R&D to close some of the gaps."

Operating in parallel means that each core can work its part of the puzzle simultaneously with other cores on a chip, greatly increasing the speed a processor operates on data. The method does not require faster clock speeds, measured in faster gigahertz, which would generate unmanageable amounts of heat to dissipate as well as current leakage.

The new method bolsters the continued relevance of Moore's Law, the 1965 observation of Intel cofounder Gordon Moore that the number of transistors placed on a single computer chip will double approximately every two years.

Another problem for the institute is to reduce the amount of power needed to run a future exascale computer.

"The electrical power needed with today's technologies would be many tens of megawatts -- a significant fraction of a power plant. A megawatt can cost as much as a million dollars a year," says Dosanjh. "We want to bring that down."

Sandia and Oak Ridge will work together on these and other problems, he says. "Although all of our efforts will be collaborative, in some areas Sandia will take the lead and Oak Ridge may lead in others, depending on who has the most expertise in a given discipline." In addition, a key component of the institute will be the involvement of industry and universities.

A spontaneous demonstration of wide interest in faster computing was evidenced in the response to an invitation-only workshop, "Memory Opportunities for High-Performing Computing," sponsored in January by the institute.

Workshop organizers planned for 25 participants but nearly 50 attended. Attendees represented the national labs, DOE, National Science Foundation, National Security Agency, Defense Advanced Research Projects Agency, and leading manufacturers of processors and supercomputing systems.

Ten years ago, people worldwide were astounded at the emergence of a teraflop supercomputer -- that would be Sandia's ASCI Red -- able in one second to perform a trillion mathematical operations.

More recently, bloggers seem stunned that a machine capable of petaflop computing -- a thousand times faster than a teraflop -- could soon break the next barrier of a thousand trillion mathematical operations a second.

Source: Sandia National Laboratories

Permalink: http://www.sflorg.com/comm_center/tech/p277_06.html

Time Stamp: 2/21/2008 at 2:04:25 PM CST

Videogame Makers Should Market to Women

Jan. 25, 2008

During the recent holiday season, consumers spent millions of dollars on videogames. While their commercial success is unquestionable, it’s amazing to think that videogames have become so successful while almost willfully excluding a sizable chunk of the population women.

Videogame developers could take some tips from Parker Brothers 100 years ago. Parker Brothers understood the most successful board games would bring the whole family together to play: male and female, parents, children and grandparents. This belief impacted how they developed and marketed their games. Recent research I conducted with the women’s game collective Ludica revealed some interesting nuances: board game covers featured intergenerational groups playing together; girls and women were as actively involved as males. Also, designers like George Parker often employed women to manufacture their products; as a result, they were often recruited to playtest his new games. This approach led to games that young and old, male and female could enjoy playing together, as well as tremendous success for Parker Brothers.

In contrast, much of the videogame industry stubbornly insists on targeting its products to the narrow audience of “hardcore gamers” comprised of predominantly high school and college-aged males. The games are largely designed by and for men. A recent study revealed that 88.5 percent of game developers are male, tested by young men and marketed to young men. And they continue to do this in the face of overwhelming evidence that shows they are missing out on a huge opportunity to move from developing niche products to developing true blockbusters.

First of all, look at the numbers: Women make up 52 percent of the U.S. population, but only 38 percent of the videogame players. Recent studies that include web-based and downloadable games have found that women over 40 spend more time on average playing games than any other group. Despite this large and interested market, female gamers are often maligned as players of exclusively casual games (short-term play, downloadable games, such as Diner Dash), a claim that is wholly unsubstantiated by any empirical research. The mainstream game industry often marginalizes these games in favor of the 40-hour or more playtime, marquee, adrenaline-infused games like Halo 3 and Grand Theft Auto produced for consoles. Imagine the opportunity if companies actually marketed to instead of against the female gamer!

Second, time and time again, titles with a strong female appeal Pac-Man and Myst, among others—have proven to be among the most commercially successful games. Pac-Man was secretly designed as a game that would appeal to women, a fact its creator, Toru Iwatani, hid from his employers. The enduring success of Pac-Man over the past 20+ years indicates that the designer was on to something. Some recent offerings, such as Dance Dance Revolution, Guitar Hero and the recently released Rock Band, follow up on this more inclusive tradition.

Third, we need look no further than the tremendous success of Nintendo’s Wii gaming console and its handheld DS. When they started developing their next generation gaming systems, Sony and Microsoft placed their bets on higher-end graphics, pushing umpteen million polygons per second, while managing not to push the boundaries of game design all that much. Nintendo decided it didn’t stand a chance in this race and its best bet was to play another game entirely.

With both the Wii and DS (as their “GameBoy” was conspicuously renamed), Nintendo has boldly gone where no game company dared to go: to new audiences. Nintendo has unabashedly made known its strategy: its competitors can have the “hardcore gamers,” they’ll take everyone else (along with some hardcore gamers too)! They are the first game company in history to have a booth at the AARP annual convention. (Baby Boomers, anyone?)

With the fastest selling handheld ever in the DS and the Wii console outselling the PlayStation 3 and Xbox 360 combined, Nintendo’s bet has paid off. Despite its lower price point, the Wii is also more profitable per unit than either of its competitors. The gaming industry has taken notice. Sony, Microsoft and Nintendo are now distributing downloadable games, popular with female gamers. Recently, Microsoft announced a new marketing and advertising campaign designed to reach casual gamers, predominantly women.

But will marketing a product designed by and for young men to a broader audience help compete for this new market into which Nintendo has tapped? Time will tell, but in the meantime, Nintendo could well position itself as the Parker Brothers of the 21st Century, finding its way into the hearts and homes of…well…everyone.

Source: Georgia Institute of Technology

Permalink: http://www.sflorg.com/comm_center/tech/p210_05.html

Time Stamp: 1/25/2008 at 11:10:29 AM CST

Sandia researchers develop integrated energy-water model for planning/management purposes

Jan. 24, 2008

First program to address fundamental resources together

Water and energy are inextricably linked. It takes large volumes of water to produce energy and significant amounts of low-cost energy to treat and distribute water.

But the planning and management of these fundamental resources have historically been done independently of one another.

A Sandia National Laboratories research team is attempting to remedy the situation by developing interactive computer modeling tools that integrate the two for planning and management purposes.

"Our model will allow energy and water producers, resource managers, regulators, and decision makers to look at the different tradeoffs of water use and energy production caused by uncertainties in population, energy demand, climate, and the economy," says Vince Tidwell, principal investigator.

Specifically, the model will help answer questions dealing with possible energy and water shortfall scenarios for particular regions; tradeoffs between alternate energy futures to meet projected shortfalls; tradeoffs between alternate water allocation schemes; economic and environmental consequences of these alternative futures; and potential consequences of alternative energy, environmental, and/or water policies.

The research is in its second year of three-year funding through Sandia's internal Laboratory Directed Research and Development (LDRD) program.

Sandia is a National Nuclear Security Administration (NNSA) laboratory.

The idea for the modeling program grew out of the Energy-Water Roadmap Development exercise conducted by Sandia and several other entities that addressed major energy- and water-related issues facing the country. The roadmap, which is under review by the Department of Energy, shows that energy and water utilities usually don't work together to resolve joint issues. This lack of coordination could lead to inefficiencies, conflict, and unnecessary stress on natural resources and the environment.

Tidwell says that currently, electrical power generation requires about 140 billion gallons of water per day, accounting for more than 40 percent of all freshwater withdrawals in the nation. That means thermoelectric generation withdrawals alone are almost equal to those for irrigated agriculture. However unlike agriculture, only a small fraction of the water withdrawn for power production is actually consumed (3.3 billion gallons per day). Nevertheless, withdrawals taken from waterways and aquifers can lead to overdraft conditions while return flows represent a source of thermal pollution.

With power demands expected to increase 30 percent by 2025, Vince questions where the water is going to come from to accommodate new power generation. Related to this is the fact that much of the growth in the US is occurring in the Southwest, a region that already has limited water.

Sandia's computer modeling initiative -- drawn on the Labs' expertise in energy, water, and optimization -- might provide some answers to these types of complex issues.

In developing the water-energy model, Tidwell says, the researchers face three problems -- the coupling of complex systems, integration of processes over disparate time and length scales, and the analysis and optimization of simulation results.

Concurrent with the energy-water modeling, the research team will put together an optimization toolbox that will assist in data analysis. Specifically, optimization will be used to help in the siting of power plants, balancing the energy portfolio (e.g., fossil, nuclear, renewables) to keep pace with growing power demands, and decisions concerning when to build the next power plant. Such decisions might consider cost, availability of water, availability of fuels, access to transmission lines, and greenhouse gas emissions.

"Users will be able to run hundreds of scenarios and see the effects in graphs and tables of their water and energy choices a year from now or decades away," Tidwell says.

The model, which will run on a standard PC, uses readily available software and provides rapid feedback.

Sandia researcher Len Malczynski, who is doing energy modeling, says the team is now compiling data to go into the program. The model will allow users to tailor their investigations to meet specific needs. For example, they can get results on energy and water scenarios at the national, state, or local levels and will be able to look at specific watersheds. This would be particularly helpful in determining water-energy trends in states like New Mexico where most of the power is generated at in-state plants but used by people from out of state.

"Energy data is provided by DOE, and water information is coming from different agencies," says Peter Kobos, who is also doing energy modeling at Sandia. "The challenge will be to have enough data to tell a story. We think we do. If not, we'll identify gaps and address them as the project progresses."

When the project is completed, the researchers expect the model will be available to water and energy utilities, regulators, and decision makers.

Source: Sandia National Laboratories

Permalink: http://www.sflorg.com/comm_center/tech/p209_004.html

Time Stamp: 1/24/2008 at 9:56:25 PM CST

Cray Introduces Next-Generation Supercomputers

Tuesday, November 6, 2007

Cray Extends Scalable MPP Architecture; Introduces Industry's First Integrated Hybrid Supercomputer With Cray XT5(TM) Family

Global supercomputer leader Cray Inc. (NASDAQ: CRAY) today launched the Cray XT5(TM) family of supercomputers, marking a significant step toward Cray's vision of adaptive supercomputing. Incorporating all the benefits of the successful Cray XT(TM) line, the Cray XT5 massively parallel processor (MPP) system includes a new compute blade that quadruples local memory capacity, doubles processor density and improves energy efficiency for a significant reduction in total cost of ownership for customers. Continuing to drive industry leadership in extreme scalability, the Cray XT5 family also includes the industry's first integrated hybrid supercomputer, the Cray XT5h(TM) system. With new programming paradigms and processor technologies, the Cray XT5h system optimizes application efficiency and helps customers tackle problems that cannot currently be solved on a single architecture system.

"The Cray XT5 family delivers superior sustained application performance from a single cabinet to massive scale while offering lower power consumption, industry-leading high density packaging, innovative cooling technologies and a fully upgradeable path from Cray XT3 and Cray XT4 systems," said president and CEO Peter Ungaro. "This new product family builds on the scalability of the Cray XT product line with the added productivity benefits of hybrid processing to deliver shorter times to solution and competitive advantages to our customers over a wider range of applications."

Cray XT5 Supercomputer

The Cray XT5 system is the world's most scalable Linux-based supercomputer. It combines unprecedented sustained application performance with exceptional manageability, lower cost of ownership and broad application and tools support to large supercomputer class machines. A number of prominent independent software vendors (ISVs) have pledged their support of the new system, including CD-adapco, Exa Corporation, Livermore Software Technology Corp. (LSTC), and Software Cradle.

In addition to supporting the current Cray XT4(TM) compute blades, a new eight-socket Cray XT5 compute blade supports the powerful dual-core and new Quad-Core AMD Opteron(TM) processors. Each dual-socket node supports up to 32 gigabytes of locally addressable memory which can be accessed at up to 25.6 gigabytes per second. Using a next-generation Cray SeaStar2+(TM) interconnect, the entire system is designed to scale and avoid performance losses associated with communication bottlenecks, memory access delays and operating system jitter.

Recognizing the growing need to reduce energy usage and control operating costs, the Cray XT5 family employs innovative packaging and technologies that reduce power and cooling requirements. Vertical cooling takes cold air from the floor with a single, high-efficiency turbine fan and efficiently cools the processors on the Cray XT5 blades. The compute blades are designed for optimal airflow with position-dependent, custom-designed heat sinks. A single cabinet can support very high density CPU configurations of 192 processor sockets or 768 CPU cores.

Easily upgradeable and expandable, existing customers can upgrade to the Cray XT5 system from the Cray XT3(TM) or Cray XT4 systems and/or add on to their existing Cray XT systems, thereby leveraging their investment over a longer life. Cray XT5 cabinets can be configured with Cray XT4 compute blades, for optimized compute-to-communication balance, or with the new high density Cray XT5 compute blades, for memory-intensive and/or compute-biased workloads. Additionally, the Linux operating environment in the Cray XT5 system enables optimal performance across a broader range of applications.

Cray XT5h Hybrid Supercomputer

Recognizing that no single architecture is ideally suited for all types of applications, Cray has developed the industry's first integrated hybrid supercomputer, the Cray XT5h system. With the Cray XT5 system as its foundation, the Cray XT5h supercomputer integrates multiple processor architectures with a complete software development environment into a single system supporting diverse workflows. The Cray XT5h system couples industry-leading scalar processing capability with high memory-bandwidth vector processing and reconfigurable co-processing using field programmable gate array (FPGA) technology, establishing a new paradigm in high performance computing (HPC).

"The stalling out of standard microprocessor speeds is starting to spark a rebirth in the HPC industry toward computers that augment microprocessors with other processor types including vector processors, GPUs, accelerators and FPGAs," said Earl Joseph, IDC program vice president, Technical Computing Systems. "Cray is developing new computer architectures to address this growing requirement with its new hybrid supercomputer. Cray has a long history and experience using multiple processors in its supercomputers including vector processors, custom I/O processors, industry standard microprocessors, multithreaded processors and FPGAs which will add value and capability to their new hybrid designs."

Supporting a variety of processing technologies, the Cray XT5h system integrates its Opteron support with a vector system and FPGAs.

Source: Cray Inc.

Permalink: http://www.sflorg.com/comm_center/tech/p150_03.html

Time Stamp: 11/6/2007 at 7:37:53 AM CST

Digital Eyes in the Sky Play Key Role in Battling Flames in Southern California

Tuesday, October 30, 2007

While staff evacuated, innovative local wireless network provided real-time images of the fires

Video and still images captured in real time have informed fire crews and local residents in the San Diego area about the location and severity of threats to life and property since fires broke out earlier this month. Accident investigators from the California Department of Forestry and Fire Protection (CAL FIRE) have requested these images that were captured by the National Science Foundation (NSF)-supported High Performance Wireless Research and Education Network (HPWREN).

Perched on mountains and bluffs overlooking the greater San Diego area, HPWREN's remote cameras have served as constant eyes in the sky, even as fires engulfed the towers on which they stood, making the situation too hazardous for human observers. These real-time cameras provide CAL FIRE and other local fire crews with a commanding view of fires as they happen, confirming fire situations on remote mountaintops saving valuable time and personnel.

Hans-Werner Braun, director of HPWREN at the University of California, San Diego, whose own family was forced to evacuate, continues to monitor the network's cameras to watch workers extinguish the fire.

"The HPWREN real-time cameras tell us what is happening before engines or chiefs can get there," says CAL FIRE Emergency Command Center Chief Tom Gardner. "They tell us clearly where to go when we are getting swamped with locals calling it in."

Chris Hinshaw, manager of the San Diego County Sheriff's Department's wireless communications office, further described the cameras' contributions as "especially helpful as they are set up with a 360-degree view from each site in small thumbnails so you can browse quickly. We were also able to estimate potential damage to our radio site by observing the fire as it moved through the facilities, and observe remedial actions such as fuel deliveries for our generators using the cameras. They are a force multiplier for us."

In addition to CAL FIRE, rural communities in the greater San Diego area created blogs that point residents to HPWREN cameras, enabling residents to check the whereabouts of blazes and acquire related information not otherwise available from traditional media.

The Lyon's Peak cameras have been especially useful for people in the Jamul community, who monitored camera footage through their neighborhood blog.

"I've heard from many Jamulians about the cameras, all with basically the same message: the cameras were what kept them sane. They were the only reliable source of information about where the fire was burning," said local resident Tom Dilatus. "It was extremely comforting to see the tiny stationary lights of what we guessed were fire engines stationed along Sierra Cielo and the dirt road to the east of our property as the flames came through last night. At least we knew that the necessary forces had been deployed."

The goal of the HPWREN project is to demonstrate and evaluate a non-commercial, prototype, high-performance, wide-area wireless network in San Diego and Riverside counties. The network includes backbone nodes on the San Diego Supercomputer Center and San Diego State campuses and a number of hard-to-reach areas in remote environments. It hosts the HPWREN servers and provides for its Internet connectivity, supporting the access to the camera images fore more than 10,000 users during the fires, at more than 10 gigabytes a day. On Oct. 25, downloads from one server alone added up to more than 70 gigabytes of data.

In addition to assisting state and local authorities with wireless networking technologies during emergencies, HPWREN is used for network analysis research and provides high-speed Internet access to field researchers in geophysics, astronomy and ecology as well as educational opportunities for rural Native American learning centers and schools.

For background information about a newly established regional command and control communications network and HPWREN's involvement, visit: http://hpwren.ucsd.edu/news/20070531/.

Image Caption 1: This shot from the real-time HPWREN-connected atop Lyons peak was taken and archived on 7:48am on Wednesday, Oct. 24, 2007, during the height of the Harris fire, which burned almost 100,000 acres in San Diego County last week.

Image Caption 2: HPWREN, when new, perched on a San Diego area mountaintop.

Image Credit 1- 2: HPWREN, funded by National Science Foundation.

Source: NSF

Permalink: http://www.sflorg.com/comm_center/tech/p145_02.html

Time Stamp: 10/30/2007 at 3:12:05 PM CST