Scientists to simulate 14 billion years of cosmic evolution

Scientists at LJMU will take major steps forward in understanding the origin and evolution of our Universe, using the UK’s largest supercomputers.

Computational astrophysicists Professor Robert Crain, Dr Andreea Font, Dr Robert Grand and Professor Ian McCarthy of the Astrophysics Research Institute will use the UK’s pre-eminent scientific supercomputing facility to simulate 14 billion years of cosmic evolution, from just after the Big Bang all the way to present day.

The ARI staff are part of a team awarded over half a billion core hours on Science and Technology Facilities Council's DiRAC network. The awarded time, the largest ever DiRAC allocation for an astrophysics project, is equivalent to more than 5,000 standard computers running for 10 years and will produce a colossal quantity of data - tens of petabytes, equivalent to 10,000 LJMU OneDrive allocations - much of which will be transferred to LJMU for analysis on the University’s centralised high performance computing facility, Prospero. 

The allocation was made to the Virgo Consortium for Cosmological Simulations, which is a collaboration between physicists in the UK, Germany, The Netherlands, Canada, the United States, and Japan. The ARI’s numerical simulations group constitutes the second largest UK node of the Virgo Consortium, and played a leading role in the DiRAC bid. The team was also awarded a similarly large allocation in 2022, securing computing time with an in-kind value of over £17 million in the last two years.

Redefining simulations

The award will enable the ARI to build on their recent successes, develop and analysing state-of-the-art computational models of the cosmos and remain at the frontier of an intensely competitive international discipline. The group is particularly interested in understanding how structure formed and evolved in the Universe, including the formation of the very first galaxies recently unveiled by the James Webb Space Telescope, and in identifying the elusive natures of dark matter and of dark energy. These two substances dominate the dynamics of the Universe but physicists still do not know what they are made of or precisely how they interact. 

Professor Robert Crain said: "High performance computing now underpins discoveries in nearly every aspect of science and technology, but perhaps nowhere so acutely as astrophysics and cosmology, because one cannot `experiment’ with the Universe. This award will enable us to develop the world’s most sophisticated `digital laboratories’ with which to advance our understanding of the complex astrophysics of galaxy evolution and the fundamental physics of the cosmic dark matter and dark energy."

A key step taken since the team’s last award is the adoption of artificial intelligence methods enabled by the University’s investment in Prospero. Crain adds: “Artificial intelligence and machine learning methods are now fundamental to both the development and the exploitation of our simulations. They enable us to explore the parameter space of possible models in a mathematically rigorous way, and reveal complex relationships between the many properties of cosmic structures that emerge from the simulations. This advance has been enabled by the addition to Prospero of nodes equipped with GPUs, which dramatically accelerate these processes.” 

Professor Ian McCarthy, Director of the ARI said: "This is a major success for our group.  It will allow us to shed light on some of the most pressing challenges in physics, including helping to uncover what is driving the accelerated expansion of the Universe. We will build on the previous generation of simulations that represent the current state-of-the-art, such as ARTEMIS, EAGLE, E-MOSAICS and BAHAMAS, and which have proven to be major international research highlights forming a major component of the ARI’s REF2021 submission.

"The award recognises the innovative work of the ARI cosmological simulations group and paves the way for exciting breakthroughs in the future. It’s particularly gratifying that significant fractions of the allocation were made in support of projects being led by the ARI’s PGRs”.

Those PGRs include Anna Durrant, who is developing simulations to examine the influence of the demographics of stellar populations on the observable properties of galaxies, and Emily Costello, who is seeking to understand the influence of black holes on the properties of the hot, diffuse gas that is bound to galaxy clusters, the most massive structures in the Universe.  PGR students Jonah Conley, Ryan Roberts and Andrea Sante are using advanced machine learning methods trained on the simulations to develop statistical models for the influence of dark matter on the growth of structure in the Universe, with a view to making predictions for forthcoming observational surveys. The students are applying the data science training they’ve received as part of the LIV.INNO Centre for Doctoral Training to their thesis projects.  

The ARI’s numerical simulations team also recently benefitted from a nearly £1 million investment from STFC awarded to Crain and McCarthy that will be used to recruit two postdoctoral researchers to work on DiRAC programme.