Intensive computer use in science

Initiatives such as Blue Waters, which cost US$ 170 million, and the Virtual School of Computational Science and Engineering were presented at the São Paulo School of Advanced Sciences.

By Luiz Paulo Juttel, Campinas

Agência FAPESP – Science guided by data: In the assessment of Yan Xu, program manager at Microsoft Research, this new paradigm has gained significance in cutting-edge scientific research.

The Internet and other computer platforms currently offer an enormous quantity of data that represents uncharted scientific areas. The “Fourth Paradigm” is the name given to scientific discoveries based on investigating pre-existing data produced by others.

Xu participated in the São Paulo School of Advanced Science on e-Science for Bioenergy Research (SPSAS e-SciBioenergy), which was held in October by the Bioethanol Science and Technology Center (CTBE) and was supported by FAPESP under the auspices of its São Paulo School of Advanced Science funding.

One trend discussed at the school was the use of complex computational systems in studies that simulate large-scale scientific phenomena by using large volumes of data.

The participants presented simulations on advances related to bird flu in the North American population and the incidence of earthquakes around the globe, among other topics.

“One major desire in the global scientific community is to have enough technology to make accurate forecasts in real time about certain phenomena and to be able to propose appropriate responses to these occurrences,” says Xu.

Another highlight of the ESPCA e-SciBioenergy was the presentation given by the director of the National Center for Supercomputing Applications (NCSA) at the University of Illinois, Thom Dunning. The NCSA manages Blue Waters, one the three largest complexes of applied computer-intensive science in the United States.

Blue Waters cost US$ 170 million, and an additional US$ 26 million is invested annually to operate the system – some US$ 13 million of which is only for electricity expenditure. The majority of this cost is funded through large federal agencies, such as the National Science Foundation (NSF) and the scientific arm of the Department of Energy (DOE).

Dunning explained that there has been growing interest in Blue Waters among private sector companies. According to the guidelines, these companies can occupy up to 5% of the system’s annual usage time.

“Companies have sought us out not only to improve their products with the help of high-performance computing but also to minimize environmental impacts. A company in the consumer products sector, for example, developed a project to redesign plastic packaging for its products. They sought something resistant while at the same time using the least amount of plastic possible. Aircraft, engine and agricultural machinery producers, among others, have also worked with us,” he said.

According to the researchers, the next step in this type of cooperation is to take intensive computer use to the large companies that are key suppliers of the production chains of these companies.

Speakers at the ESPCA e-SciBioenergy highlighted the need for concern regarding the lack of specialized labor for e-Science projects. According to these speakers, the increased number of projects in this area has led to competition for professionals with sophisticated expertise in computational systems that deal with large volumes of data and the analysis and visualization of this information.

To minimize this situation in the United States, the University of Illinois created the Virtual School of Computational Science and Engineering. The initiative has already trained more than 700 graduate students in the past few years on topics that are normally not covered in regular computing courses, such as heterogeneous computing and petascale (quadrillions of bytes) programming.

In Brazil, two important initiatives are taking shape in this area. One of these, according to Roberto Cesar Junior, a researcher at CTBE and Universidade de São Paulo (USP), is the approval of a project on e-Science methods by FAPESP’s Program to Support Centers of Excellence and CNPq.

Another project, the e-Science Research Center, which focuses on the biological, medical, human and agrarian sciences areas, has been created at USP. According to Cesar Junior, these two actions aim to increase the critical mass and knowledge networks and to install infrastructure to share e-Science services.