SUMMARY: A new supercomputer has been installed at the Rochester Institute of Technology to simulate the interactions between black holes and the evolution of galaxies. Known as the gravitySimulator, this computer will run simulations that calculate the gravitational interactions between thousands of individual stars. It can achieve a top speed of 4 Teraflops (4 trillion instructions a second), making it one of the top 100 supercomputers in the world.

View full article

What do you think about this story? Post your comments below.

I like the idea of superfast computers.

However in this case the faster computer will be able to give the wrong answers - faster.

This is because the laws of gravity (verified by Newton in our solar system) have never been shown to work - unmodified - at large galactic distances. In fact the puzzles caused by trying to explain the flat rotational velocities in spiral galaxies (reported by by Vera Rubin) leading to the need for massive dark matter must be taken as proof that Newton's laws are not accurate at large, galactic, distances.

When the gravitational constant is generalized by adding a simple term linear in distance, the conflict and dark matter disappear. Many other puzzles are solved.

I suggest that the programming of the fast supercomputer incude the generalized gravitational constant: Gn + A*r..

Details are provided at: http://inventing-solutions.com/simplified-universe.htm

Sol Aisenberg, Ph.D.

http://www.universetoday.com/forum/style_i...ges/1/icon1.gif

Hi Sol welcome to the UT forum,

We've communicated briefly about this about a year ago. I am personally skeptical about ideas in the MoND family of ideas, but open minded enough to not rule them out entirely.

I think my three biggest concerns about them are:
- There does not seem to be a constant "A" from your equation that works for all galaxies and clusters. For some galaxies the critical radius is much smaller than for others. This is especially noticable in the Low Surface Brightness Galaxies as compared to the Giant ellipticals.
- There is some evidence for self-annihilating dark matter concentrations near the center of our galaxy, in the form of anti-proton annihilation gamma rays. If CERN starts cranking out neutralinos in three years, will that invalidate MoND?
- If the scale of the gravitational force from a mass is inversely linear with distance in the long range distance, why doesn't the Universe collapse in on itself?

__________________
Forming opinions as we speak

Why not use this super computer gravity simulator to simulate the big bang theory?
If the universe just blew up from nothing and has been evolving via the fundamental laws of physics ever since, then this super computer should be able simulate the universe from the time of the big bang until what the universe currently looks like right now. Am I right or am I right?
Just put in the initial values of nothing and watch it evolve into what it is now!

You cannot fully simulate a universe. But 4 million stars is a very good start. To see what kind of feeding patterns of SMBH's are seen over billions of years is more than enough for me.

There are post-surface of last scattering situations that the new supercomputer may be good at handling.

However, from the PR itself, it's not at all clear whether the GRAPEs are optimised for Newtonian many-body interactions, or if the 'force law' that governs the interactions is programable.

If the latter, then MOND style interactions would be straight-forward to simulate (and maybe so would full GR ones). However, simulation of a radiation-dominated universe (i.e. before the surface of last scattering) would very likely require a quite different approach to the structure of the simulation, and this supercomputer would seem rather poorly suited to such.

How much can the fastest supercomputer go through a second.

How do we know if the universe will collapse or expand?

Eiter way it will take millions of years to happen - and we will not be around to observe and learn.

All we can do is speculate based upon observations.