There’s no secret that GPUs have some extreme muscle behind them, and a team of researchers at Michigan Technological University are
harnessing this power
to better understand the most complicated of real-life systems.
The project, lead by Roshan D’Souza is supercharging agent-based modeling, a powerful and computationally massive forecasting technique, with the goal of modeling complex biological systems such as the human immune response to the tuberculosis bacterium.
Mikola Lysenko, the computer science student that wrote the software demonstrated the ability of the program. A demo showing an impressive swarm of bright green immune cells surrounding and containing yellow tuberculosis bacterium was the product of millions of real-time calculations. D’Souza claims “I've been asked if we ran this on a supercomputer or if it's a movie.”
D’Souza’s only real concern is being able to do more with the technology, “We can do it much bigger,” he says. He hopes to model how a tuberculosis bacterium infection could spread from the lung to a patient’s lymphatic system, blood and vital organs.
Agent-based modeling is something that will be used to revolutionize medical research. Dr. Gary An, a surgeon specializing in trauma and critical care at Northwestern University’s Feinberg School of Medicine is pioneering its use. He’s doing so by modeling another matter of life and death, sepsis. These infections, which consist of billions of agents (including cells and bacteria), have had too complex of a model to map – until now.
While admittedly most of us will need our own supercomputer to decipher the medical jargon used to simply describe the actions of the GPU powered agent-based modeling, there’s no doubt that the results will be astonishing. And it appears that
they’re not the only ones
taking advantage of this supreme power.
Image Credit: Fastra