Illinois researcher uses XSEDE to contribute to black hole image
Charles Gammie, a professor of Physics and Astronomy at the University of Illinois, used supercomputer simulations to help confirm the first-ever captured image of a black hole.
|Black Hole GIF|
By: Boswell Hutson, National Center for Supercomputing Applications
Thanks in part to resources allocated through the Extreme Science and Engineering Discovery Environment (XSEDE) and the Blue Waters supercomputer at the National Center for Supercomputing Applications (NCSA), a massive research collaboration was able to confirm a picture of a black hole for the first time ever this month.
A research team led by Charles Gammie, a professor of Physics and Astronomy at the University of Illinois at Urbana-Champaign and a long-time XSEDE user, played a vital role in laying the groundwork in the imaging of the M87 black hole. The image was captured by the Event Horizon Telescope (EHT) collaboration, a global network of radio dishes that use short radio wavelengths to construct the equivalent of a giant interferometer, which helps to track the size and movement of black holes.
But scientists can't simply look into the sky and see a black hole, especially when none have ever been seen before. Instead, researchers must perform simulations to determine what exactly to look for, forming a blueprint based on prior observations and astrophysical processes surrounding black holes. Behind the viral image of a black hole is a massive collaboration of scientists, like Gammie, who are harnessing the power of supercomputing to interpret these groundbreaking astronomical discoveries and inform future observations.
"I had allocations on TeraGrid for aspects of this work, and then continued using NSF-funded supercomputers during the transition from TeraGrid to XSEDE," said Gammie. "So I think we've been using XSEDE since it was founded."
Through simulations on XSEDE-allocated resources and the Blue Waters system at Illinois, Gammie and his team were able to build massive astrophysical models. First, Gammie's lab used Blue Waters to investigate magnetic turbulence, one of the many mysteries shrouding black holes.
"These [turbulence] simulations gave us more confidence in the simulations that we did to model M87. This process of turbulence in the gas that's circulating around the black hole is fundamental to the entire process. We wanted to check whether we were getting that right, or whether we had to use 2x or 10x the resolution to get it done. Blue Waters enabled us to do that initial check."
With these initial simulations in hand, Gammie's team used an XSEDE allocation on the Stampede2 supercomputer, located at the Texas Advanced Computing Center (TACC), to build off the earlier work and conduct the bulk of their simulations.
"On Stampede2 we ran large hydro simulations of the gas orbiting and falling into the black hole. We also ran synthetic imaging software, developed in our group, that takes the data from these large hydro simulations and makes a relativistically consistent image from it," said Gammie. "We include the bending of light around the black hole and time delays and all of the strange things that happen around black holes."
Through these simulations, Gammie and his team provided important analysis to the imaging process, building theoretical models for researchers to use as a guide for what a potential black hole image might look like, and what informs their actions.
"When we went to interpret the EHT data on Stampede2, we used these simulations to build a huge library of simulations and synthetic images based on the simulations. We then compared the images to the EHT results to see which models worked and which didn't," said Gammie. "We ran around 60 simulations and we have in excess of 60,000 synthetic images based on those simulations."
"In particular, we found that the models in which the black hole is not spinning, do not work," Gammie continued. "Only the black holes with spin were consistent with all of the EHT data."
With the theoretical models and high-resolution simulation, Gammie and his team were able to provide clear insight for the EHT collaboration, ultimately contributing important information to the first-ever image of a black hole, an emphatic moment in human scientific history.
This work, which was published in The Astrophysical Journal Letters, used the Extreme Science and Engineering Discovery Environment (XSEDE), which is supported by National Science Foundation grant number ACI-1548562. The study, "Event Horizon Telescope, Black Holes and the Origin of Earth's Moon," is supported by XSEDE grant number AST170024.