Distributed Computing

Overview

UMIACS supports several distributed computing systems that harvest unused computing cycles from systems in UMIACS and from our partners outside the University and matches them with large scientific workloads that need centuries of computing time to be solved.

Selected Facilities

UMIACS Condor Pool

The UMIACS Condor Pool is a High Throughput Computing system built on the Condor Project from the University of Wisconsin. This system harvests resources and accepts workloads from computing clusters and workstations within the Institute. During peak periods, the system offers over 1,000 processors for serial computing jobs.

The Lattice Grid

The Lattice Grid is a federated computing system that uses the Globus toolkit and the Berkeley Open Infrastructure for Network Computing (BOINC) to match computing jobs with widely distributed resources from the Institute, from other partners in the University System of Maryland, and from a world-wide community of users who contribute computing cycles.