News from the San Diego Supercomputer Center
April 2006
Friends and Colleagues,
The advance of science, engineering, and technology is often accomplished by defining a "frontier" goal and leapfrogging forward. On the high performance computing front, the community is driving towards the development of petaflop computers, computers which can execute 10^15 operations per second. At a recent workshop, NSF described a new RFP — due in the next few months — to provide computational resources to the community which can sustain petaflop performance by 2010.
Developing petascale-level high performance architectures is one challenge on the path to a petascale; scaling applications to be able to take advantage of these extraordinarily powerful machines (many times faster than the current fastest machines on earth) will require a concerted effort over the next few years in targeting and developing petascale codes. Research communities are already beginning to put in the work that will be required to have petascale applications which can take advantage of petascale machines for breakthrough discoveries.
As part of this effort, SDSC hosted the National Science Foundation (NSF) Petascale Computing for the Geosciences Workshop in early April, with a focus on extreme application scalability. Geoscientists and computer scientists teamed at the workshop to identify target applications and plan development efforts over the next few years for petascale GEO applications. Chaired by SDSC's Allan Snavely, and funded by the Division of Atmospheric Sciences in the GEO Directorate of NSF, the workshop is a critical step in identifying and building petascale applications for geoscientists. Workshop slides can be found online. As always, we appreciate your feedback and welcome your visits.
More nuggets in May. Fran Berman and Vijay Samalam
| Predicting the State of the Ocean | |
 |  The ocean's great capacity to store heat and greenhouse gases gives it a vital role in climate change studies. Climatic trends are only one motivator for ocean study; the oceans also play a significant role in many other issues of human concern. Carrying heat, salt, nutrients, pollutants, and icebergs, ocean currents affect fisheries dynamics, shipping, offshore mining, and international policy. SDSC staff members have been working with the ECCO (Estimating the Circulation and Climate of the Ocean) Consortium to understand the state of the world's oceans, both in the past and present. To gain a complete picture of the ocean's state, ocean observations are interpolated into a highly scalable parallel simulation code MITgcm, which runs on SDSC's IBM supercomputer DataStar. Recently the team completed a series of runs -- taking 200,000 processor-hours and 4 months -- that validated the model's predictions with observations from the year 2000. The findings were unveiled at the Ocean Sciences Meeting in Hawaii in a presentation entitled "Towards an Eddy Permitting Southern Ocean State Estimate." More information on ECCO can be found on their web site and on the SDSC-ECCO collaboration. |
| SDSC Expands Archival System | |
 | SDSC currently houses about 3 petabytes (3 X 10^15 bytes) of data with a total capacity of 6 petabytes. This sounds immense, but the research and education communities are producing enormous quantities of data, and many applications utilize this data intensively. SDSC's DataCentral community collections span science, engineering, social sciences, and the arts and humanities — and our data volume is growing exponentially. In April, work began on a major archival upgrade that will triple the size of SDSC's storage capacity from 6 petabytes to 18 petabytes. The expansion in capacity comes from the addition of another tape silo (that makes 6 silos housed at SDSC) and installation of a new generation of tape drives. The upgrade will also increase bandwidth to the archival system by more than a factor of 5. SDSC's unique data cyberinfrastructure environment supports a national data repository and powerful high performance computing resources. Check the Web for more information on SDSC's production systems. |
| A Unique Look at the 1906 San Francisco Earthquake | |
 | Large-scale earthquakes of magnitude 6.0 and above can result in immense loss of life and billions of dollars in damages. Striking before dawn on April 18, 1906, a magnitude 7.9 earthquake began off the San Francisco coast and ruptured the ground for some 300 miles along the San Andreas Fault. Creating a sustained shaking that lasted more than 40 seconds, the earthquake and resulting fires left much of the Bay Area in ruins. To commemorate the 100th anniversary of this earthquake, the U.S. Geological Survey (USGS) Menlo Park asked SDSC visualization experts Steve Cutchin and Amit Chourasia to create a detailed visualization of the 1906 earthquake based on a recent simulation done for the 100th Anniversary Earthquake Conference. The resulting movies provide a compelling visual of movement of seismic waves and their intensity and took two months to generate. For more information on the San Francisco earthquake, see http://earthquake.usgs.gov/1906/. |
| SDSC Hosts Digital History for U.S. Library of Congress | |
 | The Prudkin-Gorskii images are a digitized collection of photographs documenting the Russian Empire on the eve of World War I and the coming revolution, taken by a man often referred to as the photographer to the Tsar of Russia. Documenting current culture, the Internet Archive documents modern Web content harvested from the Heritrix web crawler. Both the Prudkin-Gorskii and the Internet Archive collections are important cultural resources for present and future generations. Through a project with the Library of Congress and the UCSD Libraries, SDSC will host both collections in this unique partnership. The effort will help create a "path forward" to ensure that the Library's digital information will persist through the continual evolution of storage systems and media, operating systems, applications software, and content formats. Technology transition is a challenge for digital preservation - some of SDSC's oldest collections have already been through 6 generations of technologies. The Library of Congress collaboration will contribute significant experience and infrastructure to support cultural and scientific data preservation. For more information on the important work done by the Library of Congress in digital preservation, see www.digitalpreservation.gov. |
| Predicting Weather on the Sun | |
 | The recent solar eclipse on March 29, 2006 gave earthbound observers and scientists an opportunity to view the Sun's ghostly outer atmosphere — the corona. Normally, the corona can only be seen with very special instruments that block out the intense light from the surface of the Sun. But during a solar eclipse, the Moon (which appears about the same size as the Sun in the sky) blocks that light naturally and the beautiful corona is directly visible. The temperature of the solar corona is 2 million degrees and the resulting coronal plasma is controlled by complex solar magnetic fields. At times, "solar storms" can eject this plasma in the direction of the Earth, resulting in potentially serious disruptions in satellite operations, communications and even electrical power grids. Since our society is heavily dependent on this infrastructure, predicting the "solar " is of tremendous importance. The recent eclipse gave scientists from the Solar Physics Group at SAIC (Science Applications International Corporation) an opportunity to check their predictions of the state of the solar corona based on a computational model using observed photospheric magnetic field data. Using dedicated time on SDSC's 15.6 TeraFlops IBM supercomputer DataStar and NASA's (National Aeronautics and Space Administration) Columbia system, researchers at SAIC were able validate models of corona density, providing better predictions of the effects of coronal heating, the conduction of heat, and other factors. The calculation used over 600 processors on DataStar and took more than 4 days to complete. For more information, click here. |
