Tunnels in space-time: why scientists simulate wormholes on supercomputers

We often imagine wormholes as portals for space travel — from science fiction. But physicists see them not only as intriguing exotics, but also as a way to test the deepest laws of the universe. One of these projects, supported by the Frontiers of Science Science and Technology Guild, is located at the junction of fundamental theory and advanced computing technologies. His goal is to model in detail for the first time what happens to an "abandoned" wormhole if it loses the matter holding it open.

 

What is a wormhole and why does it collapse?

 

Wormholes are not an invention of screenwriters. They arise as one of the possible solutions to the equations of Einstein's general theory of relativity. In fact, these are hypothetical "bridges" or tunnels connecting remote regions of space-time.

 

The problem is that any such tunnel tends to snap shut instantly under its own gravity. To prevent this from happening, we need so—called exotic matter, a substance with a negative energy density that creates an "anti-gravity" effect. However, nothing like this has yet been discovered in the real universe.

 

We pose the question in a different way than most previous studies," explains Nikita Shirokov, an independent researcher with a master's degree in physics at MIPT. — Not "how to build a passable wormhole," but "what happens if it originated in quantum foam at the dawn of the universe, evolved to macroscopic scales in the era of inflation, but is now left alone, without exotic matter to keep it open." Having lost this "support", Nora is doomed. But that's exactly how she's going to die, and that's what we want to find out.

Revolution in Computing: Why Physicists are switching to GPU

 

Einstein's equations are a system of complex nonlinear differential equations that cannot be solved with pen and paper. For half a century, there has been a separate direction for solving them — numerical relativity. And today, a real tectonic shift is taking place in this area.

 

The boom in artificial intelligence has led to the rapid development of graphics processing units (GPUs). It turned out that these chips, originally created for rendering games and later for training machine learning models, are ideal for parallelizing heavy mathematical operations.

 

Previously, calculating the evolution of merging black holes on classical supercomputers with conventional CPUs could take many weeks, but modern GPU clusters with H100-level graphics cards reduce computing time by orders of magnitude. What used to be considered months are now considered days. This gives a huge scope for numerical experiments, Nikita notes.

The project uses the advanced GRTeclyn codebase (the successor to the famous GRChombo code), which is currently being adapted to the GPU. Part of the job is to teach GRTeclyn how to calculate extreme gravity on a GPU by adding the missing physical modules.

 

How to "slice" space-time: the ADM stage

 

The project is currently at the stage of active programming, first runs on clusters, and simulation analysis. The focus is on ADM production.

 

The abbreviation ADM comes from the names of the physicists Arnowitt, Deser and Mizner, who proposed an elegant mathematical trick in the middle of the 20th century. Instead of trying to describe four-dimensional space—time in its entirety, it is "sliced" into three-dimensional layers - like film frames. Each frame is a geometry of space at a certain moment in time.

At this stage, we are mathematically defining the very first "frame": the exact geometry of a wormhole without supporting matter within the framework of the ADM formalism. Then we give it a small asymmetric perturbation and force the code to calculate each next frame on the GPU. This is how we dynamically see how space itself bends, tears or collapses, says Nikita.

Is it possible to "see" the death of a wormhole in reality?

 

In fact, yes — with the help of special telescopes that catch a special type of radiation — gravitational. Such telescopes already exist — these are the LIGO, Virgo and KAGRA ground-based laser interferometers. In 2015, they first recorded gravitational waves from black hole mergers. They managed to do this only because physicists, thanks to numerical relativity, calculated in advance for them the theoretical pattern of the signal from the merger of black holes.

 

If our simulation shows that a collapsing wormhole produces a unique pattern unlike anything else, this pattern can be passed on to astronomers,— Nikita Shirokov continues. — They will load it into their algorithms and start looking for the same spikes in real data. If we prove that the burrow dissipates without radiation or the signal is indistinguishable from noise, this is also a huge result. This will give cosmologists an answer to the question of why we still do not observe relic wormholes in the modern Universe: without exotic support, they simply cannot survive in a classical vacuum.

In any case, the project is not just testing an exotic hypothesis. He is honing numerical modeling techniques and testing new approaches to supercomputer computing on the GPU. Even if wormholes remain a theoretical exotic, the tools created today will help us better understand black holes, neutron stars, and the very fabric of space-time.