This would-renowned research group works on projects that push the limits of high performance computing. In addition to their collective experience and expertise, the facilities available only at Indiana University provides this unique group the power to explore new frontiers in technology and data analytics. Big Red 200, our new artificial intelligence system, is the fist next-gen supercomputer based on HPE's Cray Shasta architecture deployed in the world. Jet Stream, another innovation led by the Indiana University faculty, adds cloud-based, on-demand computing and data analysis resources to the national cyberinfrastructure.
Push the limits of high performance computing
Subareas
ATLAS (Automatically Tuned Linear Algebra Software)
Be part of ongoing research that uses empirical tuning to optimize dense linear algebra software. Participate in the development of an empirical tuning framework available as an open source/free software package ("ATLAS").
Cost-efficient scientific machine learning
The SciSpot and TFMD (TensorFlow Molecular Dynamics) systems enable cost and time efficient execution of machine learning and scientific computing applications. Our research combines research in performance optimization, cloud computing, and distributed systems, to build practical platforms that make it easy to harness the power of cloud and HPC systems for a wide range go applications.
Efficient scientific data management
Contribute to the development of production-quality data reduction software for scientific data at scales. Explore to development of compression accelerated parallel I/O library and communication library for both distributed machine learning (ML) and HPC applications.
High performance message processing
Help in the development of a parallel software industry. Encourage the development of portable and scalable large-scale parallel applications. Contribute to the development of Photon, a high performance, remote direct memory access library for high performance computing (HPC); and Proteus, reconfigurable computing for high performance graph processing.
Modern systems & practices
Explore the development of commodity off-the-shelf (COTS) clusters and the innovations open through The Beowulf Project, which fosters the development of similar commodity off-the-shelf (COTS) clusters. Focus on game-changing computing architectures that are essential for exascale data and computer problems.