U Michigan Designs Data-Centric Supercomputer
Researchers at the University of Michigan (U-M), in collaboration with IBM,
have designed a high-performance computing (HPC) cluster with the goal
of advancing predictive modeling in computational science.
The HPC cluster, named ConFlux, is hosted at U-M's Center for Data-Driven Computational Physics and enables "large scale data-driven modeling of multiscale physical
systems," according to information on the center's site. This type of data
modeling is very challenging and requires HPC applications running on
external clusters to connect with large data sets at run time.
"The
recent acceleration in computational power and measurement resolution
has made possible the availability of extreme scale simulations and
data sets," said Karthik Duraisamy, director of U-M's Center for
Data-Driven Computational Physics, in a prepared statement. "ConFlux
allows us to bring together large scale scientific computing and
machine learning for the first time to accomplish research that was
previously impossible."
Some of the large-scale, data-driven research projects that will use ConFlux include:
- A collaboration with NASA to use cognitive techniques to simulate turbulence around aircraft and rocket engines;
- A
project for the National Institutes of Health that combines noninvasive
imaging with a physical model of blood flow to help doctors estimate
artery stiffness;
- Studying how clouds interact with atmospheric circulation in order to better understand climate science;
- Research into the origins of the universe and stellar evolution; and
- Predictions of the behavior of biologically inspired materials.
ConFlux was funded by a $2.4 million grant from the National Science Foundation and an additional $1.04 million from the University of Michigan.
IBM is providing servers and software solutions for ConFlux. Several members of the OpenPower Foundation, an open, collaborative technical community based on IBM'S Power architecture, also contributed to its development.
About the Author
Leila Meyer is a technology writer based in British Columbia. She can be reached at [email protected].