TeraShake Simulation of the ‘Big One’ Shakes California Basins

Magnitude 7.7 Earthquake Scenarios Produce Massive Data Collections at SDSC

The southern San Andreas fault in California has not experienced a major earthquake since 1680, and accumulated stress could lead to a magnitude 7.7 event. Now earthquake scientists have used the San Diego Supercomputer Center’s (SDSC) DataStar supercomputer to produce the largest and most detailed simulation yet of such earthquakes.

A key goal was to explore the response to the temblor by Southern California’s deep, sediment-filled basins. To do this, the researchers modeled a volume 600 kilometers long by 300 kilimeters wide and 80 kilometers deep, spanning all major population centers in Southern California. Dividing the volume into a grid of 1.8 billion cubes, 200 meters on a side, TeraShake produced unprecedented amounts of data—some 47 terabytes—equivalent to more than four times the printed collection of the Library of Congress.

“The TeraShake simulation is the fulfillment of a dream we’ve had for over ten years,” said scientist Jean-Bernard Minster of the Institute of Geophysics and Planetary Physics (IGPP) at UCSD. “To understand big earthquakes we need as much detail as possible, and this requires massive amounts of data from a facility with the data and computing resources of SDSC.”

Two complementary simulations were run for the same 230-kilometer stretch of the fault. A key finding was that the direction of the rupture dramatically focused the ground shaking. When the fault ruptured from north to south, the energy was focused into the Imperial Valley region in the south (fig. 1), whereas in the northward-running rupture the shaking was stronger and longer in the San Bernardino and Los Angeles basins (fig. 2). In addition to advancing basic earthquake science, the detailed simulations can eventually help structural engineers design more earthquake-resistant structures.

Led by the Southern California Earthquake Center Community Modeling Environment (SCEC/CME), participants are SDSC/UCSD, USC, ISI, IGPP/SIO, SDSU, UCSB, and CMU.

Reference: Jordan, T. H., and P. Maechling, The SCEC Community Modeling Environment—An Information Infrastructure for System-Level Earthquake Science, Seismological Research Letters, 74, 324-328, 2003.

URL: SCEC Community Modeling Environment - http://www.scec.org/cme/

NSF CA No. ACI-9619020. NPACI. PI: Francine Berman.

Decoding the Secret Language of Bees

SDSC Ensures Crucial Bee Data Kept Safe for Years to Come

Bees, along with humans, occupy a very exclusive place in the world of animals that have developed abstract language for describing their surroundings. Why is it that some bee species have evolved this capability? With the help of the San Diego Supercomputer Center (SDSC) and international collaborators, biologist James C. Nieh of UCSD is pursuing research to answer this question.

A key tool that the scientists rely on is digital video, which they use to record bee communication. The researchers have discovered that some bees have the ability to use sounds to encode information about food location. This ability can prevent other bee species from intercepting the information, and Nieh believes that such eavesdropping may have helped drive the development of sophisticated bee languages as anti-espionage encoded communication to transmit food source information to nestmates inside the hive.

However, using digital video is data-intensive, and requires the researchers to store and provide access to massive amounts of information. SDSC is providing critical cyberinfrastructure in the form of data expertise and technology that helps the scientists store, manage, and analyze their data. For each bee species, the scientists record 1.2 terabytes of digital video annually. The archive, which the researchers expect to grow to 30 terabytes or more, is stored in SDSC’s six petabyte archive (one petabyte is one million gigabytes).

In addition to storing the digital video, SDSC data and networking infrastructure helps widely-separated collaborating labs in Mexico, Brazil, Panama, and San Diego have efficient distributed access to the data as they analyze millions of video frames of bee behavior. In addition to supporting current science, SDSC provides a home to keep irreplaceable scientific collections available long into the future for research and education.

URL: James Nieh lab - http://www-biology.ucsd.edu/faculty/nieh.html

NSF CA No. ACI-9619020. NPACI. PI: Francine Berman.

Improving Biomass Conversion to Ethanol for Renewable Energy

SDSC Helps Enhance CHARMM Community Code to Study Key Reaction

The National Renewable Energy Laboratory (NREL) is striving to develop efficient large-scale conversion of biomass into ethanol to provide a clean-burning and renewable fuel source. This can have benefits from reducing dependence on fossil fuels and imported oil to protecting the climate. A key bottleneck in making this process economically viable is the slow breakdown of cellulose by the enzyme cellulase, and scientists want to understand this process at the molecular level so that they can target further research to speed up this important reaction.

To explore the intricate molecular dynamics of this process, the researchers have used the CHARMM (Chemistry at HARvard Molecular Mechanics) code, a versatile community code for simulating biological reactions. But the size of new simulations needed is so large—more than one million atoms—and the simulation times so long, more than 5,000 time steps for the 10 or more nanosecond simulation, that they exceed the current capabilities of CHARMM.

Now researchers at the San Diego Supercomputer Center (SDSC) are working with colleagues at NREL, Cornell, the Scripps Research Institute, and the Colorado School of Mines to enhance CHARMM so that the simulations can scale up to millions atoms and run on hundreds of processors on today’s largest supercomputers, including SDSC’s DataStar and the TeraGrid. This will make it feasible to simulate this key reaction.

The research is enabling the largest simulations ever of an important scientific problem that will yield economic and environmental benefits, and in addition, the improvements to the CHARMM code will continue to be available for the scientific community to use on a wide range of other problems.

Reference: Sheehan, J. & Himmel, M. (1999). Enzymes, energy, and the environment: A strategic perspective on the US Department of Energy's Research and Development Activities for Bioethanol. Biotechnology Progress, 15, 817-827.

URL: http://www.nrel.gov/biomass/

NSF CA No. ACI-9619020. NPACI. PI: Francine Berman.

In the Beginning: Exploring the Early Universe with Enzo

Combined SDSC Data and Computational Resources Enable Massive Simulations

In one of the most complex scientific simulations ever performed of the formation of structure in the universe, astrophysicists have used the San Diego Supercomputer Center’s (SDSC) DataStar supercomputer to calculate the evolution of the distribution of visible matter—galaxies, quasars, and gas clouds—starting just 30 million years after the Big Bang and running for more than three billions years.

The comparison of the simulations and telescopic observations of distant objects provides a “reality check” on astrophysicists’ theories of the origin of the universe and the formation of stars and galaxies, and the simulation results are serving as a source of data for research papers by astrophysicists. “Computational science—using computer simulations to explore the consequences of physical laws—has become an essential part of modern research,” said astrophysicist Robert Harkness of SDSC. “Astronomers, after all, don’t have any other way to bring galaxies into the laboratory and perform experiments on them.”

The scientists ran their “Enzo” cosmology code on DataStar in a simulation volume 1,024 cells on a side (more than a billion cells in all) modeling a volume of space some 248 million light-years on a side. Complementary simulations took an extremely detailed look at 0.5 percent of the full volume, with far shorter time steps and finer spatial detail.

Not only are the simulations computationally intensive, they generated many tens of terabytes of data, presenting the researchers with major challenges in data management, analysis, and visualization. “SDSC is the only place in the world at this time where this simulation can be done and the scientific content analyzed, because of SDSC’s investment in data management technology,” said project leader Michael Norman, an astrophysicist at UCSD.

Reference: Tytler, David et al, Cosmological parameters ?8, the baryon density ??, the vacuum energy density ??, the Hubble constant and the UV background intensity from a calibrated measurement of HI Lyman ? absorption at z = 1.9, Astrophysical Journal, 617:1-28, 2004.

URL: UCSD Center for Astrophysics and Space Science, CASS -- http://casswww.ucsd.edu/

NSF CA No. ACI-9619020. NPACI. PI: Francine Berman.

SDSC’s Rocks Software Wins Three Awards at Supercomputing 2004

Rocks Software Makes Cluster Management for Scientific Computing Easy

As scientific applications have grown more demanding and commodity clusters less expensive, the need for software that makes clusters easy to build and use has grown. To meet this need, Rocks, funded by the NSF and developed at the San Diego Supercomputer Center (SDSC), provides powerful and user-friendly cluster management.

Recognizing the enthusiastic community adoption of Rocks software, HPCWire, the publication for high performance computing, presented the Rocks team at the recent SC2004 conference with two Reader’s Choice Awards and an Editor’s Choice Award for software innovation.

Rocks is an easy-to-use software package that gives the user of a cluster computer the ability to quickly and easily build the Linux system software suite, install new software on the processors, and manage the system configuration.

The Rocks user base includes five clusters on the Top500 list of the world’s largest supercomputers, with several hundred Rocks clusters deployed around the globe. The Rocks Register, where users voluntarily register their clusters, shows Rocks powering an aggregate of nearly 70 teraflops of peak computing,

The current version of the Rocks Clustering Distribution is 3.3.0, which supports all commodity processors, x86, Opteron, Itanium, and now EM64T Xeon. This release includes two new Rolls, for Visualization and Infiniband. Rolls are optional extensions to the base distribution that support specific cluster types or enhanced functionality.

The Rocks development community includes SDSC, Scalable Systems in Singapore, the University of Tromso in Norway, Kasetsart University in Thailand, and the Korea Institute of Science and Technology Information (KISTI). More information on Rocks, including documentation and complete access to the source code, can be found on the project’s homepage.

URL: Rocks Homepage - http://www.rocksclusters.org.

NSF CA No. ACI-9619020. NPACI. PI: Francine Berman.