Visualization of Large Scientific Datasets - Analysis of Numerical Simulation Data and Astronomical Surveys Catalogues

In the context of our project COAST (for Computational Astrophysics), a program of massively parallel numerical simulations in astrophysics involving astrophysicists and software engineers, we have developed visualization tools to analyse the massive amount of data produced in these simulations. We present in this paper the SDvision code capabilities with examples of visualization of cosmology and astrophysical simulations realized with hydrodynamics codes, and more results in other domains of physics, like plasma or particles physics. Recently, the SDvision 3D visualization software has been improved to cope with the analysis of astronomical surveys catalogues, databases of multiple data products including redshifts, peculiar velocities, reconstructed density and velocity fields. On the basis of the various visualization techniques offered by the SDvision software, that rely on multicore computing and OpenGL hardware acceleration, we have created maps displaying the structure of the Local Universe where the most prominent features such as voids, clusters of galaxies, filaments and walls, are identified and named.