See epic computer sim re-create the universe
In a project of galactic proportions, scientists plug the laws of physics, matter, dark matter, and dark energy into a supercomputer network to simulate the development of our universe, starting just after the Big Bang.
What if you could take a snapshot of yourself shortly after birth, convert all the things that make you you at that early age into data, plug that information into a computer simulation, and then let that simulation run for a few decades using all we think we know right now about how humans develop into adulthood? Would the computer-simulated you look like the real adult you or more like Max Headroom?
Scientists from MIT, Harvard, and institutions of great nerdiness in Europe have run a similar simulation using a network of supercomputers, but rather than simulate the development of a single person, they brought up the entire universe from when it was just a wee little entirety of all existence to the far more inconceivable everything that it is today.
This ultra-complex computer model uses a kind of "universe in a cube" approach. The model represents an area of space that is almost 350 million light years across, which the scientists say is large enough to be representative of the whole universe, but it also allows users to "zoom in" to check out individual galaxies, right down to a scale as small as just 1,000 light years across.
The researchers created over 100,000 lines of code simulating the laws of physics, our understanding of ordinary matter, dark matter, and dark energy and input a snapshot of our universe as it would have appeared just 12 million years after the Big Bang. From that point, they let the simulation run forward over the next 13 billion years -- our universe is believed to be about 13.8 billion years old.
A team effort
Running that entire simulation took computing power of nearly galactic proportions. Had it been run on a single top-of-the-line consumer desktop, it would have taken thousands of actual years to run. Fortunately, a networked collaboration of supercomputers at Harvard, the Texas Advanced Computing Center, and European facilities allowed it to run much more quickly, over a span of a few months.
In the video below you can see some of the simulation in action and watch virtual galaxies forming around dark matter while black holes often stir the quasi-cosmic pot.
In the end, the simulation modeled 41,416 galaxies, just a fraction of the actual universe, but what's amazing is just how close that small sample is to the real thing, including the creation within the model of galaxies just like our own.
"For the past two decades, cosmologists have been unable to produce galaxies like the Milky Way in their simulations," David Spergel, a professor of astronomy at Princeton University, said in a release.
But this latest model managed to produce spiral galaxies just like the Milky Way.
"Some galaxies are more elliptical and some are more like the Milky Way, [spiral] disc-type galaxies," explained Mark Vogelsberger of MIT, the first author of a new paper in Nature that describes the project. "There is a certain ratio in the universe. We get the ratio right. That was not achieved before."
Of course, the simulation does have some flaws.
"Low-mass galaxies just form too early in the simulation,"Vogelsberger added. This means that many galaxies appear much older than their analogs would be in the real universe.
A solid virtual foundation
Still, if you take a shot from the Hubble telescope of the actual observable universe and put it next to a virtual shot from a virtual Hubble within the simulation, the similarities are clear (see above). And keep in mind, the simulation was guided by nothing more than basic data from our primordial universe and the laws of our universe as astrophysicists understand them today.
So the good news is that we seem to understand our universe fairly well, at least on a galactic scale 1,000 light years across.
"But there's some room for improvement in the future," Vogelsberger said in an interview. "It's not the successes of the model that puts us forward, it's definitely the failures because then we can learn what is wrong in our understanding and try to come up with better theories."