Playstations settle black hole related controversy
In September 2007, Dr. Gaurav Khanna of UMass Dartmouth clustered eight (8) Sony Playstation 3 gaming consoles together and demonstrated that one could obtain supercomputer-level performance at an extremely low cost. This project generated strong interest in the scientific community and also in the popular media about the possibility of making use of such alternative hardware for serious numerical computation. Since then, the PS3 cluster has been expanded (with an upgrade of additional eight PS3s), and has been used for research projects in the area of black hole astrophysics.
Rotating black holes are the most common form of black holes that are known to exist, astrophysically. In fact, the centers of most galaxies harbor very large such black holes. Our own Milky Way galaxy has in its core a "supermassive" black hole whose mass is comparable to four million times that of the Sun. Astrophysical black holes are described by only two parameters: their mass and their spin angular momentum. When a black hole first forms, or when it is perturbed (e.g., by swallowing some matter) it radiates away the additional parameters, and settles down to a quiescent state that is fully described only by its mass and spin. The question of exactly how fast it settles has been the subject of a heated and long controversy.
In a 2008 research article, that has recently been published in the IOP journal "Classical and Quantum Gravity", Dr. Lior Burko of the University of Alabama in Huntsville and Dr. Khanna settle this controversy. This research article puts that open issue to rest by performing very detailed and accurate numerical simulations using the PS3 cluster. Settling this controversial issue will help our overall theoretical understanding of rotating black holes and have an impact on the relevant area of black hole astrophysics. These results are in agreement with independent, but somewhat restricted ("special case") results obtained recently by two other groups.
This work required a very large amount of numerical computation (mainly due to the very high accuracy and length of the simulations required) and for that reason would have been very expensive to do on rented supercomputer time. The PS3 cluster was therefore indispensable for the success of this research project, whose computational needs matched the cluster's capabilities extremely well.
Drs. Burko and Khanna are continuing to use this PS3 cluster for topical problems in black hole astrophysics and gravitational wave physics, doing the kind of computations that the PS3 cluster is well suited for.
The full research article can be found at:
Author: "Dr. Gaurav Khanna"