PI: Tom Peterka, MCS

Description: We propose to study the layering of HPC programming models beneath big data programming models. Specifically, we are researching how to build a software environment with a Spark user interface (Java and Scala) that can run on a supercomputer, cluster, or cloud with a DIY (MPI, C++) back end for executing data-intensive communication patterns.

Other Collaborators: University of  Carlos III of Madrid, Spain

Testbed: Sage

Schedule: One Year