· Desk editor: Adam Whitmore · · Single source

Researchers simulate Milky Way stars using three supercomputers

A team led by University of Tokyo researchers presented a record-resolution Milky Way simulation at the SC 2026 conference in Chicago from November 15 to 20.

New model tracks individual stars

The research group developed an AI-accelerated method to model the entire galaxy down to single stars. This approach combines power from three different supercomputer systems using mixed CPU and GPU hardware. The simulation handles rapid changes in dense gas regions that previously required too many calculation steps.

AI predicts supernova gas evolution

Postdoctoral researcher Keiya Hirashima from RIKEN created the AI model that predicts how gas evolves around supernovae. Directly calculating this process normally requires time steps of around 100 years for each event. The new model allows the main simulation to skip these smaller calculations and run much faster.

Efficiency gains for galaxy study

The researchers state this method makes the simulation around 20 times faster than direct calculation methods. Professor Michiko Fujii noted that finer detail increases computational cost significantly for large galaxies like the Milky Way. The team expects better simulations as overall computational power continues to increase in the future.

Reported by one outlet

Only one outlet has published this. Nothing here has been checked against a second report, so read it as that outlet's account and follow the link below for the original.

Reported by

1 independent outlet. Headline as published. Links open the original report.