A workshop dedicated to the Theoretical Virtual Observatory will take place at IAP on April 5-6th.
The goal is to bring together experts of the Virtual Observatory and theoreticians who would like to make results of their simulations (e.g. databases or catalogs) or numerical codes available to the worldwild astronomical community.
This "grand challenge" project features the largest N body simulation ever performed: we have simulated the 13.7 Gyr long evolution of 4096^3 dark matter particles in a 2 Gpc/h periodic box, running the RAMSES code for a couple of month on the 6144 processors of the new BULL supercomputer of the CEA supercomputing center (Centre de Calcul Recherche et Technologie).
The initial displacement field was computed on a 4096^3 base grid using the MPgrafic package. We use as initial conditions parameters the best fit values inferred from the 3 years observations of the WMAP satellite. The 70 billions particles were evolved using the Particle Mesh scheme of the RAMSES code on an adaptively refined grid (AMR) with more than 140 billions cells. We use more than 10 000 fine time steps at the deepest refinement level and 737 time step at the base level. Each of the 70 billions cells of the base grid was recursively refined up to 6 additional levels of refinement. We reached a formal resolution of 262144 cells in each direction (roughly 7 kpc/h comoving).
For the first time, we have performed a simulation of half the observable universe, with enough resolution to describe a Milky Way-like galaxy with more than 100 dark matter particles.
Our goal is to generate a Full Sky mock catalog with a realistic galaxy distribution up to redshift 1, as well as a deeper catalog with 500 square degrees up to redshift 7. In this way, we are definitely approaching the scale of the cosmological horizon, which is the hard limit of the observable universe (hence the name of the project).
The following people in the Horizon Project have spent nights and days monitoring and post-processing this huge simulation: D. Aubert, S. Colombi, J. Devriendt, P. Ocvirk, C. Pichon, S. Prunet, R. Teyssier.
We describe now in more details the results of this simulation.
Here is a movie describing the simulation results at the final epoch.
For higher resolution animations, please contact us. A spectacular zoomable unfolded image is also available here. It represents an oblic slice through the simulation at z=0 16x16 Gpc/h accross.
A mollweise representation (a projection of the sphere which attempts to display all the sphere on a plane) of the simulation at redshift 1 is shown here:
It assumes that the observer views all the night sky arround him but is only sensitive to the large scale structures of the universe at a distance of roughly 5 million light years. Another partial projection is shown there:
A thick slice of the simulation at z=0. is shown here:
And a thiner slice
. A VERY LARGE (97 Mega) image in 8192x8192 pixels is available here. Note that this image corresponds only to one level of refinement (another 5 levels are available). A composite image displaying all scales probed at the initial resolution of the simulation (without refinement).
The outer region corresponds to a view of the universe on scales of 16h^-1Gpc: it is generated by unfolding the simulation while cuting a slice obliquely through the cube in order to preserve the continuity of the field (thanks to the periodicity) See also below. The intermediate region corresponds to a slice of 2h^-1Gpc, while the inner region is at the original resolution of the initial conditions. RAMSES has refined 6 times over the course of the run from that resolution. A redshift slice through the Horizon-4pi universe is shown here
while a 3D view of its filaments is shown here:
A VERY LARGE (62 Mega) image of the unfolded universe (16x16 Mpc/h) in 8192x8192 pixels is available here. Note that this image is generated from a () downgraded version of the initial cube. Another view of this cube is shown here:
while the skeleton of the structures within this cube is shown there:
Scientific Rationnal: Full sky weak lensing
A map of kappa (the weak lensing signal) is archived here (IT’S QUITE LARGE!). This image is a projection onto the sky the density of the dark matter.