BlueGene/L breaks its own record
On March 23, 2005, Administrator Linton F. Brooks of the National Nuclear Security Administration (NNSA) announced that the BlueGene/L supercomputer developed through the Advanced Simulation and Computing Program for NNSA’s Stockpile Stewardship efforts has performed 135.3 teraops (trillion operations per second) on the industry standard LINPACK benchmark. Installed in the Laboratory’s new Terascale Simulation Facility, the machine broke its previous record as the fastest supercomputer in the world. This performance was achieved with only half the BlueGene/L machine. The remaining half will be installed this summer. Last November, just one-quarter of BlueGene/L topped the Top500 List of the world’s supercomputers.
“BlueGene/L will address vital challenges critical to ensuring the safety and reliability of the nation’s aging nuclear weapon stockpile,” says Brooks. “This supercomputer provides an essential resource to the weapons complex, allowing us to address time-urgent and mission-critical scientific issues that require such specialized computational capabilities.”
Results of scientific importance have already been attained with only one-quarter of BlueGene/L. Livermore scientists for the first time have performed 16-million-atom molecular dynamics simulations to resolve the key physical effects for successfully modeling pressure-induced, rapid resolidification in tantalum.
“BlueGene/L allows us to address computationally taxing stockpile science issues at very low cost,” says Livermore Director Michael Anastasio. “Effective and relatively inexpensive supercomputers of this nature will open doors for scientists across the country.”
Contact: Dona Crawford (925) 422-1985 (email@example.com).
Mystery of ice planets unraveled
The conditions in the core of icy, large planets such as Neptune and Uranus are extreme, with pressures of hundred of thousands or millions of times greater than that on Earth and temperatures of thousands of degrees. Under similar conditions, Laboratory scientists have discovered a “superionic” phase of water—neither ice nor liquid—in which the oxygen atoms remain virtually stationary while the hydrogen atoms are extraordinarily mobile.
“This shows how extreme pressures and temperatures can completely transform water from a molecular system into a ‘salt’ composed of mobile protons and stationary oxygen ions,” says Alex Goncharov, a Livermore chemist and lead author of a paper that was published on April 1, 2005, in Physical Review Letters. “This phase of water is of profound importance not only to planetary science but also to geoscience and fundamental chemistry.”
To re-create the extreme pressure required to force water into the superionic state, Goncharov’s team used a diamond anvil cell, which squeezes the water between two diamonds, creating a pressure 470,000 times greater than Earth’s atmospheric pressure. The researchers used a laser to heat the water to the intense temperature of 1,500 kelvins. At this pressure and temperature, the water went superionic.
In addition to laboratory experiments, the team used computer models to predict the atoms’ behavior, which showed that observed changes in the optical spectra were consistent with a superionic phase. The simulations were performed using Thunder, one of Livermore’s terascale computers.
The observations and models have larger implications for the makeup of the universe. There could be more yet unobserved superionic water in the universe than water in solid or liquid forms.
Contact: Alex Goncharov (925) 422-5976 (firstname.lastname@example.org).
Study offers new clues to gene deserts
Scientists studying human chromosomes 2 and 4 have uncovered new clues to gene deserts—long stretches of DNA in the human and other mammalian genomes that contain no protein-coding genes. In a paper published April 7, 2005, in Nature, Livermore bioinformatics scientist Ivan Ovcharenko and his colleagues revealed that a particular feature of two large gene deserts has persisted through hundreds of millions of years of evolution, even though the DNA sequence within the deserts has changed considerably over time. The deserts are located on either side of two closely related genes active in the human brain and heart.
The fact that the “architecture” of the deserts has persisted over long evolutionary periods could be significant. “We were surprised to see how old the deserts were,” says Ovcharenko. “They appeared at a time point that follows the divergence of fish from Ciona intestinalis (a sea squirt) but precedes the divergence of mammals from fish. They have no obvious function, but for some reason, they were preserved at that location in the genome. This makes us wonder what’s special about these particular gene deserts.” Ovcharenko says the information obtained from the analysis of chromosomes 2 and 4 will lead to further study “to see why some genes in the human genome tend to have gene deserts in their neighborhoods.”
The Nature paper’s lead author was LaDeana Hillier at the Genome Sequencing Center (GSC) at Washington University School of Medicine in St. Louis. GSC led the international team that sequenced and analyzed the two chromosomes.
See the feature article in this issue of S&TR “Mining Genomes” for more work by Ovcharenko on gene deserts.
Contact: Ivan Ovcharenko (925) 422-5035 (email@example.com).
Better estimates of the solar nebula’s lifetime
By studying some of the oldest objects in the universe, Laboratory physicist Ian Hutcheon and colleagues from the University of Hawaii at Manoa, the Tokyo Institute of Technology, and the Smithsonian Institution have found important clues about the lifetime of the solar nebula—the mass of dust and gas that eventually formed the solar system. In their project, funded by Livermore’s Laboratory Directed Research and Development Program, the researchers measured the oxygen and magnesium content of chondrules and calcium aluminum–rich inclusions (CAIs), both components of the primitive meteorite Allende. Their results indicate that the oxygen in the solar nebula evolved rapidly over a span of 2 million years.
According to the team’s measurements, CAIs are enriched with 4 percent more of the isotope oxygen-16 (16O) than is found on Earth. (Different isotopes of an element vary in the number of neutrons in the nucleus of each atom.) This small amount of 16O enrichment is a signature of the oldest objects in the solar system.
The team’s results indicate that CAIs formed in an oxygen-rich environment about 4.567 billion years ago. Chondrules formed in an oxygen setting much like that on Earth and date to 4.565 billion years ago or less. Previous estimates of the solar nebula’s lifetime ranged from less than 1 million years to 10 million years. Says Hutcheon, “Refining the lifetime of the solar nebula is significant in terms of understanding how our solar system formed.” The team’s results appeared in the April 21, 2005, issue of Nature.
Contact: Ian Hutcheon (925) 422-4481 (firstname.lastname@example.org).