In a groundbreaking feat of computational science, the world’s second fastest supercomputer, known as Frontier, has completed the largest and most intricate simulation of the universe to date. Previously holding the title of the fastest machine until earlier this month, Frontier operates at Oak Ridge National Laboratory and is designed to push the boundaries of exascale computing, boasting a phenomenal capacity of up to 1.1 exaFLOPS—meaning it can handle 1.1 quintillion floating-point operations each second. This monstrous machine is powered by an impressive array of 9,472 AMD central processing units (CPUs) and 37,888 AMD graphics processing units (GPUs), making it a pinnacle of technological achievement.

The simulation, part of a project led by the U.S. Department of Energy's Argonne National Laboratory, employed the Hardware/Hybrid Accelerated Cosmology Code (HACC), a sophisticated code that models the evolution of the cosmos. This tool has been in development for about 15 years and has adapted to leverage the capabilities of the fastest supercomputers available at any time.

Previously, HACC was utilized on less powerful petascale supercomputers. One notable project showcased how it modeled various cosmological scenarios on the Summit supercomputer, which led the pack from November 2018 to June 2020. These simulations were creatively named after iconic planets from the Star Trek universe, such as the Qo'nos simulation, which explored the conventional model of cosmology, and the Ferenginar simulation, delving into the impacts of variable dark energy over time. The results have significant implications, suggesting that changes in dark energy could strengthen galaxy clustering in the early universe.

But Frontier takes this exploration to an entirely new level. Unlike its predecessor simulations that focused solely on gravitational effects, Frontier combines various fundamental forces and effects—what researchers like Salman Habib refer to as "cosmological hydrodynamics simulations." These factors include not just dark matter and dark energy, but also baryonic matter, which comprises everything familiar to us—including stars, black holes, and galaxies.

To achieve this, the ExaSky project was initiated, with a whopping $1.8 billion investment aimed at enabling exascale simulations that reflect true astrophysical conditions. The initial goal was for HACC to run at least 50 times faster on Frontier compared to its performance on Titan, the then-fastest supercomputer in 2012. However, in a stunning demonstration, HACC achieved nearly 300 times the speed, underscoring the phenomenal capabilities of Frontier.

The results of these cutting-edge simulations will be made publicly accessible to the astronomical community, providing valuable insights into the workings of dark matter and dark energy, and facilitating new questions regarding models of gravity, including Modified Newtonian Dynamics (MOND). Researchers will be able to correlate these simulation outcomes with real-life astronomical observations, determining the validity of various cosmological models in light of new evidence.

This remarkable technological advancement raises a tantalizing question: As computational power continues to surge, how far could our simulations of the universe evolve? Will we eventually reach a stage where we can predict cosmic phenomena with pinpoint accuracy? The possibilities are as limitless as the universe itself, inviting both awe and profound reflection on humanity’s quest to understand our place in the cosmos.