But it’s not just weight gain that poses a risk. People who are underweight also have an elevated risk of dementia, unlike people who are normal weight or overweight.

US researchers carried out a detailed review of 10 international studies published since 1995, covering just over 37,000 people, including 2,534 with various forms of dementia. Subjects were aged between 40 and 80 years when the studies started, with follow-up periods ranging from three to 36 years.

The review, which included studies from the USA, France, Finland, Sweden and Japan, also included a sophisticated meta-analysis of seven of the studies, published between 2003 and 2007 with a follow-up period of at least five years.

All kinds of dementia were included, with specific reference to Alzheimer’s Disease and to vascular dementia — where areas of the brain stop functioning because the blood vessels that supply them are damaged by conditions such as high blood pressure or heart disease.

“Our meta-analysis showed that obesity increased the relative risk of dementia, for both sexes, by an average of 42 percent when compared with normal weight” says Dr Youfa Wang, Associate Professor of International Health and Epidemiology at Johns Hopkins Bloomberg School of Public Health, Baltimore.

“And being underweight increased the risk by 36 percent.

“But when we looked specifically at Alzheimer’s Disease, the increased risk posed by obesity was 80 percent. The increased risk for people with vascular dementia was 73 percent.

“The risks were greater in studies where sufferers developed Alzheimer’s Disease or vascular dementia before the age of 60 or in studies with follow-up periods of more than 10 years.

“We also found that obesity was more likely to be a risk factor for women when it came to developing Alzheimer’s Disease and for men when it came to vascular dementia.”

The authors estimate that 12 percent of the dementia risk in the study population could be attributed to obesity, with this rising to just over 21 percent in patients with Alzheimer’s Disease.

It’s estimated that up to 10 percent of people aged 65 or more suffer from some form of dementia and two-thirds of those have Alzheimer’s Disease.

“There has been controversy about the links between obesity and dementia for a number of years, but previous findings have been mixed and inconclusive” says Dr Wang.

“The advantage of carrying out a meta-analysis is that it provides researchers with access to a large number of study subjects and it is possible to iron out the inconsistencies and come to overarching conclusions.

“Our detailed analysis clearly shows a U-shaped relationship between weight and dementia, with people who are obese or underweight facing a greater risk.

“We believe that our results show that reducing the prevalence of obesity is a promising strategy for preventing the progression of normal ageing into Alzheimer’s Disease.”

[Annette Whibley @ Wiley-Blackwell]

Berkeley Lab has signed a collaboration agreement with Tensilica, Inc. to explore the use of Tensilica’s Xtensa processor cores as the basic building blocks in a massively parallel system design. Tensilica’s Xtensa processor is about 400 times more efficient in floating point operations per watt than the conventional server processor chip shown here.

In a paper published in the May issue of the International Journal of High Performance Computing Applications, Michael Wehner and Lenny Oliker of Berkeley Lab’s Computational Research Division, and John Shalf of the National Energy Research Scientific Computing Center (NERSC) lay out the benefit of a new class of supercomputers for modeling climate conditions and understanding climate change. Using the embedded microprocessor technology used in cell phones, iPods, toaster ovens and most other modern day electronic conveniences, they propose designing a cost-effective machine for running these models and improving climate predictions.

In April, Berkeley Lab signed a collaboration agreement with Tensilica, Inc. to explore such new design concepts for energy-efficient high-performance scientific computer systems. The joint effort is focused on novel processor and systems architectures using large numbers of small processor cores, connected together with optimized links, and tuned to the requirements of highly-parallel applications such as climate modeling.

Understanding how human activity is changing global climate is one of the great scientific challenges of our time. Scientists have tackled this issue by developing climate models that use the historical data of factors that shape the earth’s climate, such as rainfall, hurricanes, sea surface temperatures and carbon dioxide in the atmosphere. One of the greatest challenges in creating these models, however, is to develop accurate cloud simulations.

Although cloud systems have been included in climate models in the past, they lack the details that could improve the accuracy of climate predictions. Wehner, Oliker and Shalf set out to establish a practical estimate for building a supercomputer capable of creating climate models at 1-kilometer (km) scale. A cloud system model at the 1-km scale would provide rich details that are not available from existing models.

To develop a 1-km cloud model, scientists would need a supercomputer that is 1,000 times more powerful than what is available today, the researchers say. But building a supercomputer powerful enough to tackle this problem is a huge challenge.

Historically, supercomputer makers build larger and more powerful systems by increasing the number of conventional microprocessors — usually the same kinds of microprocessors used to build personal computers. Although feasible for building computers large enough to solve many scientific problems, using this approach to build a system capable of modeling clouds at a 1-km scale would cost about $1 billion. The system also would require 200 megawatts of electricity to operate, enough energy to power a small city of 100,000 residents.

In their paper, Towards Ultra-High Resolution models of Climate and Weather, the researchers present a radical alternative that would cost less to build and require less electricity to operate. They conclude that a supercomputer using about 20 million embedded microprocessors would deliver the results and cost $75 million to construct. This “climate computer” would consume less than 4 megawatts of power and achieve a peak performance of 200 petaflops.

“Without such a paradigm shift, power will ultimately limit the scale and performance of future supercomputing systems, and therefore fail to meet the demanding computational needs of important scientific challenges like the climate modeling,” Shalf said.

The researchers arrive at their findings by extrapolating performance data from the Community Atmospheric Model (CAM). CAM, developed at the National Center for Atmospheric Research in Boulder, Colorado, is a series of global atmosphere models commonly used by weather and climate researchers.

The “climate computer” is not merely a concept. Wehner, Oliker and Shalf, along with researchers from UC Berkeley, are working with scientists from Colorado State University to build a prototype system in order to run a new global atmospheric model developed at Colorado State.

“What we have demonstrated is that in the exascale computing regime, it makes more sense to target machine design for specific applications,” Wehner said. “It will be impractical from a cost and power perspective to build general-purpose machines like today’s supercomputers.”

Under the agreement with Tensilica, the team will use Tensilica’s Xtensa LX extensible processor cores as the basic building blocks in a massively parallel system design. Each processor will dissipate a few hundred milliwatts of power, yet deliver billions of floating point operations per second and be programmable using standard programming languages and tools. This equates to an order-of-magnitude improvement in floating point operations per watt, compared to conventional desktop and server processor chips. The small size and low power of these processors allows tight integration at the chip, board and rack level and scaling to millions of processors within a power budget of a few megawatts.

Berkeley Lab is a U.S. Department of Energy national laboratory located in Berkeley, California. It conducts unclassified scientific research and is managed by the University of California. Visit our Website at www.lbl.gov.

[Ucilia Wang @ DOE/Lawrence Berkeley National Laboratory]

The method, published in the May 4 issue of Nature Nanotechnology, enables more precise shaping of microchip components than what is possible with current technology. More precise component shapes could help manufacturers build smaller and better microchips, the key to more powerful computers and other devices.

“We are able to achieve a precision and improvement far beyond what was previously thought achievable,” said electrical engineer Stephen Chou, the Joseph C. Elgin Professor of Engineering, who developed the method along with graduate student Qiangfei Xia. Chou’s lab has previously pioneered a number of innovative chip making techniques, including a revolutionary method for making nanometer-scale patterns using imprinting.

Microchips work best when the structures fabricated on them are straight, thin and tall. Rough edges and other defects can degrade or even ruin chip performance in most applications. In integrated circuits, for instance, such flaws could cause current to leak and voltage to fluctuate. In optic devices, they could interfere with the transmission of light. In biological devices, they could impede the flow of DNA and other biomaterials.

“These chip defects pose serious roadblocks to future advances in many industries,” Chou said.

To deal with this problem, researchers try to improve the process used to make the microchips. However, Chou said such an approach works only to a point; eventually chip makers will run up against fundamental physical limits of current manufacturing techniques. In particular, the electrons and photons that are used like chisels to carve out the microscopic features on a chip always have some random behavior. This effect becomes pronounced at very small scales and limits the accuracy of component shapes.

“What we propose instead is a paradigm shift: Rather than struggle to improve fabrication methods, we could simply fix the defects after fabrication,” said Chou. ???And fixing the defects could be automatic — a process of self-perfection.???

Chou’s method, termed Self-Perfection by Liquefaction (SPEL), achieves this by melting the structures on a chip momentarily, and guiding the resulting flow of liquid so that it re-solidifies into the desired shapes. This is possible because natural forces acting on the molten structures, such as surface tension — the force that allows some insects to walk on water — smooth the structures into geometrically more accurate shapes. Lines, for instance, become straighter, and dots become rounder.

Simple melting by direct heating has previously been shown to smooth out the defects in plastic structures. This process can’t be applied to a microchip, for two reasons. First, the key structures on a chip are not made of plastic, which melts at temperatures close to the boiling point of water, but from semiconductors and metals, which have much higher melting points. Heating the chip to such temperatures would melt not just the structures, but nearly everything else on the chip. Secondly, the melting process would widen the structures and round off their top and side surfaces, all of which would be detrimental to the chip.

Chou’s team overcame the first obstacle by using a light pulse from so-called excimer laser, similar to those used in laser eye surgery, because it heats only a very thin surface layer of a material and causes no damage to the structures underneath. The researchers carefully designed the pulse so that it would melt only semiconductor and metal structures, and not damage other parts of the chip. The structures need to be melted for only a fraction of a millionth of a second, because molten metal and semiconductors can flow as easily as water and have high surface tension, which allows them to change shapes very quickly.

To overcome the second obstacle, Chou’s team placed a plate on top of the melting structures to guide the flow of liquid. The plate prevents a molten structure from widening, and keeps its top flat and sides vertical, Chou said. In one experiment, it made the edges of 70 nanometer-wide chromium lines more than five times smoother. The resulting line smoothness was far more precise than what semiconductor researchers believe to be attainable with existing technology.

The conventional approach to fixing chip defects is to measure the exact shape of each defect, and provide a correction precisely tailored to it — a slow and expensive process, Chou said. In contrast, Chou’s guided melting process fixes all defects on a chip in a single quick and inexpensive step. “Regardless of the shape of each defect, it always gets fixed precisely and with no need for individual shape measurement or tailored correction,” Chou said.

One of the big surprises from this work is observed when the guiding plate is placed not in direct contact with the molten structures, but at a distance above it. In this situation, the liquid material from the structures rises up and reaches the plate by itself, causing line structures to become taller and narrower — both highly desirable outcomes from a chip design perspective.

“The authors demonstrate improved edge roughness and dramatically altered aspect ratios in nanoscale features,” said Donald Tennant, director of operations at the NanoScale Science and Technology Facility at Cornell University. The techniques “may be a way forward when nanofabricators bump up against the limits of lithography and pattern transfer,” he said.

Next, Chou’s group plans to demonstrate this technique on large (8-inch) wafers. Several leading semiconductor manufacturers have expressed keen interest in the technique, Chou said.

[Steven Schultz @ Princeton University Engineering School]