High Performance I/O

Many scientific problems make use of very large data sets. Examples include land cover dynamics, submarine structural acoustics, analysis of earth observation data, computational biology, computational quantum chemistry, seismic data processing etc. The goals of our research are to determine the I/O requirements of such applications and to develop compile-time and run-time techniques to optimize their performance on multiprocessor architectures with multiple disks or disk arrays. Currently, the major driving applications for this work are the University of Maryland's Land Cover Dynamics Grand Challenge project, submarine structural acoustics and analysis of the data generated by the Earth Observation System project which is a part of NASA's Mission to Planet Earth. This research is being carried out in collaboration with the Scalable I/O consortium and CRPC.


Parallelization and characterization of data-intensive applicationsthe goal of this project is to find out it takes to achieve good I/O performance for real-life data-intensive tasks on parallel machines suitable for I/O.

Interprocedural compiler analysis for overlapping large I/O operations with computation the goal of this project is to schedule large I/O operations which are beyond the capacity of standard operating system prefetch and write-behind.

Scheduling and resource management for I/O intensive tasks on peer-to-peer systems the goal of this project is to develop techniques for achieving high performance for I/O intensive tasks on peer-to-peer systems.

Data placement techniques for multidimensional datasets on large disk farms the goal of this project is to develop data placement techniques for multidimensional datasets both across multiple disks and within individual disks to minimize query response time.

Associated Projects

NSF/ARPA Grand Challenge project on Land Cover Dynamics this project focuses on employing high performance computing for applications in remote sensing, specifically applications in land cover dynamics. Understanding land cover dynamics is one of the most important challenges in the study of global change. Research involves developing scalable and portable programs for a variety of image and map data processing applications, eventually ntegrated with new models for parallel I/O of large scale images and maps.

Publications and Talks

An Interprocedural Framework for Placement of Asynchronous I/O Operations, Gagan Agarwal, Anurag Acharya, Joel Saltz. ICS'96

Study of Scalable Declustering Algorithms for Parallel Grid Files Bongki Moon, Anurag Acharya, Joel Saltz IPPS'96

Tuning the Performance of I/O-Intensive Parallel Applications Anurag Acharya, Mustafa Uysal, Robert Bennett, Assaf Mendelson, Michael Beynon, Jeff Hollingsworth, Joel Saltz, Alan Sussman. IOPADS'96

Jovian: A Framework for Optimizing Parallel I/O Robert Bennett, Kelvin Bryant, Alan Sussman, Raja Das, Joel Saltz, Proceedings of the 1994 Scalable Parallel Libraries Conference

Slides from the Grand Challenge Meeting, July 1994


Scalable I/O Consortium



Maryland Applications for Measurement and Benchmarking of I/O for Parallel Computers
I/O request traces from four non-scientific and three scientific applications.




Related Sites

Last Updated:  03/01/99