Hans De Sterck , Department of Applied Mathematics, University of Waterloo
Rob Knight , Department of Molecular, Cellular, and Developmental Biology, University of Colorado at Boulder
Rob Markel , Scientific Computing Division, National Center for Atmospheric Research

New: You can now download our August 2006 'Limited Experimental Release 0.1' of the TaskSpaces software [zip file].

Note that this limited experimental release is not for the faint-hearted. It is a snapshot of the research code at the end of the project, and is provided as is. The code is released in conjunction with our bioinformatics application paper and a book chapter on our software framework for parallel bioinformatics on computational grids:

• Rob Knight, Hans De Sterck, Rob Markel, Sandra Smit, Alexander Oshmyansky, and Michael Yarus,
  `Abundance of correctly folded RNA motifs in sequence space, calculated on computational grids',
  Nucleic Acids Research 33, 5924-5935, 2005. [html] [pdf]


• Hans De Sterck, Rob Markel, and Rob Knight,
  `TaskSpaces: A Software Framework for Parallel Bioinformatics on Computational Grids',
  Chapter in textbook `Parallel Computing in Bioinformatics and Computational Biology', A. Zomaya, editor, John Wiley and Sons, 651-669, 2006. [pdf]

Project description:

TaskSpaces is a lightweight platform-independent grid computing framework for scientific computing applications.

The Taskspaces framework is characterized by two major design choices: decentralization provided by an underlying tuple space concept, and object orientation and platform independence realized by implementation in Java.

TaskSpaces can be used in taskfarming mode for problems that do not require interprocess communication, but if the application requires so, the framework can also handle interprocess communication in a scalable way using communication of serialized Java objects over sockets.

TaskSpaces has been demonstrated to scale well on large grids composed of supercomputers at NCSA, SDSC, and other supercomputer centers. First applications have been explored in the fields of iterative linear solvers and Lattice Boltzmann fluid dynamics simulations.

We are currently applying the TaskSpaces framework to a large parallel bioinformatics problem. We use virtual experiments on computational grids composed of the world's fastest supercomputers to find the smallest pool of random RNA molecules that contains enough catalytic motifs for starting a primitive metabolism. This may establish one of the missing links in the chain of events that lead to the origin of life.

• ACM SAC 2003 paper describing the TaskSpaces framework.

H. De Sterck, R.S. Markel, T. Pohl, and U. Ruede,
`A lightweight Java Taskspaces framework for scientific computing on computational grids',
Proceedings of the ACM Symposium on Applied Computing, Track on Parallel and Distributed Systems and Networking, 1024--1030, 2003. [pdf]

• Seminar in the CU Boulder Bioinformatics Supergroup describing the TaskSpaces framework applied to a parallel bioinformatics RNA folding problem.

H. De Sterck, R.S. Markel, and R. Knight,
`A Software Framework for Parallel Bioinformatics on Computational Grids'.
Bioinformatics Supergroup Seminar, University of Colorado at Boulder, 3 November 2003. [pdf]

• Poster at the 2003 RNA Society Meeting describing the parallel bioinformatics RNA folding problem and preliminary results.

R.D. Knight, H. De Sterck, R.S. Markel, C. Lozupone, S. Changayil, I. Majerfeld, and M. Yarus,
'Finding correctly folded active RNA motifs', Poster presentation at the RNA Society Meeting, Vienna, Austria, July 1-6, 2003. [pdf]

Created by Hans De Sterck.
Department of Applied Mathematics, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada.
Phone: 1-519-888-4567 ext 7550, Fax: 1-519-746-4319, E-mail: hdesterck@math.uwaterloo.ca.
Office: MC 5016. campus map