Roadrunner Remains HPC Speed Champ

24 Jun 2009:

For a record-setting tenth consecutive time, an IBM system holds the number one position in the ranking of the world's most powerful supercomputers. The IBM computer built for the "roadrunner project" at Los Alamos National Lab -- the first in the world to operate at speeds faster than one quadrillion calculations per second (petaflop) -- remains the world speed champion.

The closely watched TOP500 list, issued twice a year, both confers bragging rights on research institutions and manufacturers and serves as a valuable tool for tracking trends in supercomputer performance and architectures. The latest list reflects changes from November 2008 to June 2009.

Holding onto the No. 1 spot with 1.105 petaflop/s (quadrillions of floating point operations per second) is the Roadrunner system at DOE’s Los Alamos National Laboratory (LANL) which was built by IBM and in June 2008 became the first system ever to break the petaflop/s Linpack barrier. It still is one of the most energy efficient systems on the TOP500.

Maintaining its hold on second place is the Cray XT5 Jaguar system installed at the DOE’s Oak Ridge National Laboratory. Jaguar reached 1.059 petaflop/s shortly after its installation but due to its heavy workload no further measurements were possible.

But in third place, a new contender has emerged-- a new IBM BlueGene/P system called JUGENE and installed at the Forschungszentrum Juelich (FZJ) in Germany. It achieved 825.5 teraflop/s (trillions of floating point operations per second) on the Linpack benchmarks and has a theoretical peak performance of just above 1 petaflop/s. FZJ is also home to the new No. 10 system. Called JUROPA, it is built from Bull Novascale and Sun SunBlade x6048 servers and achieved 274.8 Tflop/s.

The two systems in Germany are the only non-U.S.-based systems in the latest TOP10 list. There are only two other new entries in the TOP10. At No. 6 is a new Cray XT5 system called Kraken and installed at the National Institute for Computational Sciences at the University of Tennessee with a Linpack performance of 463.3 Tflop/s, making it the most powerful university-based system. The other new entry, at No. 9 with 415.7 Tflop/s, is a new IBM BlueGene/P system called Dawn installed at DOE’s Lawrence Livermore National Laboratory.

Another notable system is the Chinese-built Dawning 5000A at the Shanghai Supercomputer Center at No 15. It is the largest system which can be operated with the Windows HPC 2008 operating system.

The U.S. is clearly the leading consumer of HPC systems with 291 of the 500 systems (unchanged from 291). The European share (145 systems – down from 151) is settling down after having risen for some time, but is still substantially larger than the Asian share (49 systems – up from 47).

Furthermore, IBM declared its intent to break the exaflop barrier, and announced that it had created a research 'collaboratory' in Dublin, in partnership with the Industrial Development Agency (IDA) of Ireland, which is focused on both achieving exascale computing and making it useful to business. An exaflop is a million trillion calculations per second, which is 1000 times faster than today's petaflop-class systems.

Having ushered in the petaflop era a year ago, IBM has established a Research collaboratory in Dublin, Ireland, in collaboration with the IDA, focused on achieving exascale computing and making it beneficial for businesses with technologies like stream computing to analyze massive amounts of real-time data. This is the first collaboratory that IBM has announced, and the company intends to create more around the world.

"It's an honor to hold the record for the world's most powerful computer, but what is critical is building supercomputers that help advance the global economy and society at large," said David Turek, vice president, IBM Deep Computing. "IBM was the first to break the petaflop barrier and we will continue to apply lessons learned as we march toward the exaflop barrier."

An IBM collaboratory is a laboratory where IBM Researchers co-locate with a university, government, or commercial partner to share skills, assets, and resources to achieve a common research goal.

IBM Researchers are already at work with government and academic leaders to develop exascale systems that will help solve the complex business and scientific problems of the future. This research collaboratory will enable IBM supercomputing and multidisciplinary experts to work directly with University researchers from Trinity College Dublin, Tyndall National Institute in Cork, National University of Ireland Galway, University College Cork and IRCSET, the Irish Research Council for Science, Engineering and Technology to develop computing architectures and technologies that can overcome current limitations -- such as space and energy consumption -- of dealing with the massive volumes of real-time data and analysis.

The technical research will explore innovative ways of using new memory architectures, interconnecting technologies and fabric structures, and will evaluate business applications that would benefit from an exascale streaming platform.

While high performance computing today primarily focuses on scientific applications in areas such as physics or medicine, the exascale research in Dublin will also focus on how these new powerful computing systems can be applied to solving complex business problems. The research will include both technical and applications research. For example, the application research for exascale computing will study financial services using real-time, intelligent analysis of a company's valuation developed from business models using data from investor profiles, live market trading and RSS news feeds.

"IBM led the industry in breaking the petaflop barrier last year," continued Turek. "Developing exascale systems challenge space and energy limitations, requiring extremely sophisticated systems management and application software that can take advantage of this computational capability. This new collaboratory is already at work solving some of these issues."

As future computing platforms are expected to produce orders of magnitude more power dissipation, researchers believe that efficiently cooling these large systems will be one of the most important factors to next generation development. Making computing systems and data centers energy-efficient is a staggering undertaking.

In fact, up to 50% of an average air-cooled data center's carbon footprint or energy consumption today is not caused by computing but by powering the necessary cooling systems to keep the processors from overheating -- a situation that is far from optimal when looking at energy efficiency from a holistic perspective. IBM has numerous leading edge research projects underway that are addressing these "energy aware" hurdles.

Just today, IBM and the Swiss Federal Institute of Technology Zurich unveiled plans to build a first-of-a-kind water-cooled supercomputer that will directly repurpose excess heat for the university buildings. The innovative system, is expected to decrease the carbon footprint of the system by up to 85% and estimated to save up to 30 tons of CO2 per year, compared to a similar system using today's cooling technologies.

IBM provides a broad portfolio of systems, storage and software technology to the supercomputing market, more than any other vendor. The company's innovative HPC solutions have created a new scientific force for tackling the world's grand challenges around climate science, the hunt for new sources of energy, creating new gene-based medicines, and have made significant contributions to basic scientific inquiry in physics and biology.