Narrow Results

Rapid advances in cloud computing have made the vision of utility computing a near-reality, but only in certain domains. For science and engineering parallel or distributed applications, on-demand access to resources within grids and clouds is hampered by two major factors: communication performance and paradigm mismatch issues. We propose a framework for addressing the latter aspect via software adaptations that attempt to reconcile model and interface differences between application needs and resource platforms. Such matching can greatly enhance flexibility in choice of execution platforms — a key characteristic of utility computing — even though they may not be a natural fit or may incur some performance loss. Our design philosophy, middleware components, and experiences from a cross-paradigm experiment are described.

Finding a parallel architecture adapted to a given class of algorithms is a central problem for architects. This paper presents a methodology to realize it, and provides an illustration using image analysis. First, we show a set of common basic operations that can be used to solve most image analysis problems. Then these movements are translated to fit some natural communications in a given architecture. The considered data movements (global operations on connected pixel sets) can express a large class of algorithms. Their implementation on exemplary massively parallel architectures (arrays, hypercubes, pyramids) is discussed.

Finding a parallel architecture adapted to a given class of algorithms is a central problem for architects. This paper presents a methodology to realize it, and provides an illustration using image analysis. First, we show a set of common basic operations that can be used to solve most image analysis problems. Then these movements are translated to fit some natural communications in a given architecture. The considered data movements (global operations on connected pixel sets) can express a large class of algorithms. Their implementation on exemplary massively parallel architectures (arrays, hypercubes, pyramids) is discussed.