A growing trend in developing large and complex applications on today's Teraflop scale computers is to integrate stand-alone and/or semi-independent program components into a comprehensive simulation package. One example is the Community Climate System Model which consists of atmosphere, ocean, land-surface and sea-ice components. Each component is semi-independent and has been developed at a different institution. We study how this multi-component, multi-executable application can run effectively on distributed memory architectures. For the first time, we clearly identify five effective execution modes and develop the MPH library to support application development utilizing these modes. MPH performs component-name registration, resource allocation and initial component handshaking in a flexible way.

In solving Partial Differential Equations, such as the Barotropic equations in ocean models, on Distributed Memory Computers, finite difference methods are commonly used. Most often, processor subdomain boundaries must be updated at each time step. This boundary update process involves many messages of small sizes, therefore large communication overhead. Here we propose a new approach which expands the ghost cell layers and thus updates boundaries much less frequently ---reducing total message volume and grouping small messages into bigger ones. Together with a technique for eliminating diagonal communications, the method speedup communication substantially, up to 170 percent. We explain the method and implementation in details, provide systematic timing results and performance analysis on Cray T3E and IBM SP.

Community Climate System Model (CCSM) is currently a multi-executable system based on the Multi-Program Multi-Data (MPMD) mechanism. Each component is compiled into a separate executable. MPMD is normally cumbersome in usage and vendor support is sometimes limited or completely unavailable, such as on BlueGene/L. Also smaller groups and institutions would like to run CCSM locally, rather than relying on large computer centers. So, single-executable CCSM is under request. We are developing a multi-executable and single-executable coexisting version of CCSM. In single-executable, each component is organized as a subroutine, which is called from a master program. Different components run simultaneously. This is accomplished by redesigning the top level CCSM structures using the Multi-Program Handshaking (MPH) library.

Many current large and complex HPC applications are based on semi-independent program components developed by different groups or for different purposes. On distributed memory parallel supercomputers, how to perform component-name registration and initialize communications between independent components are among the first critical steps in establishing a distributed multi-component environment. Here we describe MPH, a multi-component handshaking library that resolves these tasks in a convenient and consistent way. MPH uses MPI for high performance and supports many PVM functionality. It supports two major parallel integration mechanism: multi-component multi-executable (MCME) and multi-component single-executable (MCME). It is a simple, easy-to-use module for developing practical codes, or as basis for larger software tools/frameworks.

We introduce a novel application of Support Vector Machines (SVMs) to the problem of identifying potential supernovae using photometric and geometric features computed from astronomical imagery. The challenges of this supervised learning application are significant: 1) noisy and corrupt imagery resulting in high levels of feature uncertainty,2) features with heavy-tailed, peaked distributions,3) extremely imbalanced and overlapping positiveand negative data sets, and 4) the need to reach high positive classification rates, i.e. to find all potential supernovae, while reducing the burdensome workload of manually examining false positives. High accuracy is achieved via asign-preserving, shifted log transform applied to features with peaked, heavy-tailed distributions. The imbalanced data problem is handled by oversampling positive examples,selectively sampling misclassified negative examples,and iteratively training multiple SVMs for improved supernovarecognition on unseen test data. We present crossvalidation results and demonstrate the impact on a largescale supernova survey that currently uses the SVM decision value to rank-order 600,000 potential supernovae each night.