Monday  April 3rd

HP Component Frameworks and Solver Interoperability: Two Years Later

Organizer and chairman: Robert Clay, SNL

Two years ago we previewed the beginnings of several efforts involving
component architectures and standards-based integration; namely, the
CCA and ESI standards efforts.  The critical issues are reviewed in
the context of the current efforts, and calling out what we've learned
in the preceding two years.

Industry, government, and university interest in this area has grown
steadily. There are far more participants, and it's core part of the ASCI
strategy for software interoperability.  Two years ago we were a bunch
of researchers exploring what seemed a promising technology.  Now, we
have built this into our core, long-term software strategy, and we are
actively building components and component-based frameworks.  Indeed, we
are practically in the midst of a revolution.

The purpose of this workshop is to discuss issues pertaining to HP
component architecture, frameworks, and applications.


+++++++++++++++++++++++++++++++++++++++++++++++++++++++

Tuesday April 4

Multi-level preconditioning techniques

Organizer: Wei-Pai Tang (Univ Waterloo)

Talks by:
Yousef Saad*
Henk van der Vorst*
Charles Tong* and Ray Tuminaro
Robert Bridson* and Wei-Pai Tang	

* is the speaker


More and more computational experience indicates that a good preconditioner
holds the key to an effective  iterative solver. However,  
many effective conventional preconditioning techniques (such as ILU,
approximate inverse, etc..) still suffer from lack of 
scalability, when the size of the problems grows.

Multi-level or multi-resolution technique offer promises when they are
combined with these conventional techniques. In this workshop, the leading
experts in this areas will discuss their new results and ideas for a
powerful preconditioner.


+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++


Wednesday April 5

Optimization Workshop

Organizers: Tim Kelley (NCSU)and Homer Walker (WPI) 

The workshop on iterative methods for optimization will include two
aspects of the field. Large scale optimization problems having thousands
or millions of variables require new methods for linear and nonlinear
equations and preconditioning as well as new algorithms for the
optimization itself. Problems with multiple local minima,
nonsmooth or discontinuous optimization landscapes, and nonlinear
constraints whose violation can only be detected when the function
fails to return a value require new optimization technology and raise
basic questions about iterative methods such as termination and
error estimation.