next up previous contents index
Next: FLY-FLASH An interface for Up: Research Previous: Collaboration with industries and   Contents   Index

Computational technologies for astrophysics

INAF Researchers: V. Antonuccio, U. Becciani
PhD Students: D. Ferro, A. Romeo
Fellows: M. Comparato, A. Calanducci
Technical Staff: A. Costa

The development of parallel computing systems and of High Performance Computing (HPC) techniques has made possible the realization of computationally demanding problems.

This type of investigations and the execution of large simulations has enhanced the presence of more problems. The management of large amounts of simulated data and their effective interpretation has required the development of new instruments to meets specific requirements. Such requirements are often complex and they must take into account the complexity of data.

Our Research Unit has given solution to all these requirements working on the development of a tool of Scientific Visualization, called AstroMD, that allows a preliminary data analysis of a sub-sample of data. The tool runs on Personal Computer under either Linux or Windows OS. Using the techniques of the Virtual Reality and with a developed tool of analysis, the user can obtain all the information to get a deeper analysis of data and of the entire dynamics. This final task is often carried out through the use of codes that make use of HPC techniques.

Until today, astrophysical and cosmological simulation, data analysis and data storage have been carried out mainly with the use of HPC systems often localized in Centers of Excellence and Supercomputing centers. Even though these systems will continue to give their meaningful contribution in the next years, it seems obvious that the development of the technologies based on the Computational Grids, will give the effective answer to the increasing requirement of the computational power, data storage and database, data mining and scientific visualization.

Within this framework the INAF-Catania Astrophysical Observatory has undertaken a line of research oriented to the development of technologies based on the Computational Grid, through some main activities. The Italian Ministry of Research (MIUR) has funded a PRIN project "Data Grid for astrophysical research: porting of scientific applications" aimed at the development of the Italian Astrophysical Computational Grid. The INAF Astrophysical Observatory partecipates in this project.

In this project a relevant effort is done to ease the use of complex simulation codes based on complementary methods and that of complex computational tools (parallel platforms with different architectures, heterogeneous platforms, visualiation tools).

A relevant asset of this project is constituted by the availability of computational codes based on different numerical approaches and of different (in terms of architectures) computational platforms; this will foster the spread of the different scientific experiences in the whole community which will certainly receive important benefits.

The specific program of the Research Unit will have the main goal to set up the Data Grid testbed site as a node of the Italian Grid for Astrophysics following the international standards in this field.

The research unit will develope both software for general purpose and concerning cosmological simulations that can be executed on the grid, the use of AstroMD as an analysis tool and visualization of data distributed on the grid and the use of the AstroComp portal developed for to the building grid. All the developed tools will be at the disposal of the scientific community.

Recently the Regional Government of Sicily using funds from the EU has financed the project "TriGrid Virtual Laboratory", with 3 Million euros (the global cost of the project being 3.75 Million euros). Using this funds a regional computational grid will be realized with more than 1.2 Million euros used to buy CPUs for Linux Cluster and HPC systems, that will be also used for computational researches in astrophysics.

In the following we report the main results of the Research Unit obtained in the last year.



Subsections
next up previous contents index
Next: FLY-FLASH An interface for Up: Research Previous: Collaboration with industries and   Contents   Index
Innocenza Busa' 2005-11-14