GPUs Help Solve Mystery of Gas Giant Formation

by Isha Salian

Jupiter. We know it as a gas giant, the largest planet in the solar system, fifth from the Sun. But until recently, astronomers couldn’t explain how it got there. Now, a team of researchers have used GPUs to reveal a breakthrough in our understanding.

Scientists believe gas planets form through a gradual buildup of ice and dust into a solid core, or protoplanet. If that solid core grows fast enough to 10 times Earth’s size, it also starts to accumulate large quantities of gas on its way to becoming a gas giant.

There’s just one problem: physics calculations suggest that the gravitational effects caused by the gas disk in which planets form, or protoplanetary disk, should push planets into a death spiral towards their star. Based on those predictions, the universe should contain lots of rocky, super-sized Earths, and fewer gas giants.

Compare the theory with observations, however, and it doesn’t add up. NASA’s Kepler mission to discover Earth-like planets orbiting stars beyond our solar system has found a higher proportion of gas giants than theory accounted for.

Protoplanet
A gas giant is born: Gas planets form by accumulating ice and dust around a solid protoplanet.

GPUs Aid Gas Planet Understanding

United by a common expertise in planet-disk interactions, four researchers from three countries recently set out to close this gap between theory and observation. Frédéric Masset and Gloria Koenigsberger from the National Autonomous University of Mexico; Pablo Benítez-Llambay from the University of Córdoba, Argentina; and Judit Szulágyi from the Observatoire de la Côte d’Azur, France, took on the challenge of modelling the physics between protoplanets and the protoplanetary disk.

Running these simulations is no simple feat. They demand immense computing power, capable of modeling the hydrodynamics of the gaseous nebula surrounding an ever growing and moving celestial body, and also of computing the transfer of radiation within this nebula. Running all their models would have taken up a million CPU hours. That’s why the research team chose a cluster of NVIDIA Tesla K20 GPU accelerators instead.

Using NVIDIA CUDA helped the researchers develop a publicly available code, called FARGO3D, which accounts for all the hydrodynamics and radiation transfer processes. Powered by the NVIDIA Kepler GPU architecture, the Tesla K20 cluster could run this code on each of the authors’ models within hours — gaining a 40x speedup on one GPU versus one CPU core.

By Jove, They’ve Got It!

The team found that as an early Jupiter and other new planets accumulate mass, they release heat. Due to the flow of gas in the protoplanetary disk, the heat is unequally distributed around the planet. The difference in temperature — and therefore density — causes the planet to move further away from its star, instead of closer.

While prior research had found evidence for outward planetary migration, inward migration was predicted too frequently to account for observational statistics. With the team’s findings, published in Nature earlier this year, the numbers at long last may add up.

The asymmetric heating of gas around these new planets explains their trajectory away from their star, not towards it. The team’s work helps account for the many gas giants NASA has observed, and represents an important step in demystifying gas planet formation.