he CBPF Computing Group aims to work in research, technological development and applications in the following areas:
• Scientific Instrumentation
• Signals and Images Processing
• Distributed Computing for Physics
• Network Systems
It is also the group’s goal to participate in technological innovation in collaboration with other research institutions and companies. Furthermore, the Computing Group is also responsible for technical services and the specialized training of human resources in interdisciplinary areas (such as physics and engineering) fostering the development of new professionals with technological innovative skills.
Scientific instrumentation: Since its foundation, CBPF has had a strong and important activity in the development of scientific instruments. This development has been the basis of technical activities in CBPF. Currently, the group is developing a widely used tool in scientific laboratories, a Lock-in Amplifier. This tool is based on Digital Signal Processors (DSP). It was designed to operate at high frequencies (2MHz) and can be applied in the study of advanced materials, such as those with technological potential, as high critical temperature superconductors, magnetic materials, manganites etc. The Lock-in was the result of a Master's thesis in the Scientific Instrumentation Program of CBPF in 2008.
Image Segmentation and Reconstruction: The group works on the development of image analysis techniques for applications in nanotechnology and magnetism, Optimized Reconstruction of signals in physical experiments, techniques based on new statistical physics theories, pattern recognition using artificial neural networks, spatial transformations and statistical methods. An example of image processing application in physics is the identification of gravitational arcs in astronomical images which are being held at the DES-Brazil Project (Dark Energy Survey). Another instance is the development of image processing techniques that allow microscopic processes analysis which occur on the surface of a material during the nanostructuring process. As a result of this, a Brazilian patent was required for the latter one.
Distributed Computing: The group works also in the NextComp Project. In this project we develop molecular dynamics simulations of long-range interactions systems which can be scalable to thousands of processors in high performance computer systems. The new algorithm was developed in object oriented language, CHARM++. In 2007, a simulation was performed with 1000 processors in the NCSA (National Center for Supercomputing Applications - USA) computing cluster. Currently, in partnership with the State University of Rio de Janeiro State - UERJ, new computer codes have been developed, in CHARM++, to simulate an X-RAY tomography by using parallel programming in high-performance computing systems. The images reconstruction calculations have been performed on the cluster of the Supercomputing Center of the Federal University of Rio Grande do Sul (CESUP / UFRGS).
Network Systems: This area is divided into two axes: R&D and specialized services. The former’s has two main interest; the first is the networks structures characterization, such as the Internet topology, its development and relationships, Complex Networks and applications of non-extensive statistical mechanics. The second concerns the characterization of the mechanisms related to the scientific citations impact.
In the latter axis, the group is responsible for the operation of the Rio de Janeiro Academic Computing Network, the Internet of the Science and Technological institutions in the state of Rio de Janeiro, a special project of FAPERJ (The State of Rio de Janeiro Research Foundation). In 2008, the Computing Group coordinated the Technical Project of the new academic network of Rio de Janeiro, named Redecomep-Rio. This project is led by the National Research and Education Network (RNP, an institution of the Brazilian Science and Technology Ministry) and by FAPERJ having various government (federal, state and municipal) institutional partners as well as private ones. The Redecomep initiative is concerned with the ever-increasing demands for greater capacity for academic institutions wide bandwidth communication. The Metropolitan Network of Rio de Janeiro will be the largest in Latin America, about 140 Gbps of total capacity and over 200 km long. In 2009, the project aims for the installation of equipment and optical fibers with DWDM technology (Dense Wavelength Division Multiplexing) with the capability of data transfer at up to 1.9Tbps in a central 10Gbps backbone.
Technological Innovation: In recent years, the group also helped with the implementation of the Center for Technological Innovation (NIT-Rio), a project aiming to bring the findings of CBPF research to the productive sector, making the institution part of the technological innovation process and encouraging the development of products and services protected by law. The NIT of CBPF is developed in partnership with National Laboratory for Scientific Computing / LNCC and National Observatory / ON. In 2008, for this project, CBPF was one of the national best classified institutions in the first phase of the project Knowledge Management in Public Centers of Research and Development in Brazil.
More information - Internet pages:
• Rede-Rio de Computadores: http://www.rederio.br
• Redecomep: http://www.redecomep.rnp.br
• Lab. Proc. Sinais e Imagens: http://www.cbpf.br/cat/lpdsi
• Coordenação de Atividades Técnicas: http://www.cbpf.br/cat