The International Journal of High Performance Computing Applications

“Grid” computing has emerged as an important new field, distinguished from conventional distributed computing by its focus on large-scale resource sharing, innovative applications, and, in some cases, high performance orientation. In this article, the authors define this new field. First, they review the “Grid problem,” which is defined as flexible, secure, coordinated resource sharing among dynamic collections of individuals, institutions, and resources—what is referred to as virtual organizations. In such settings, unique authentication, authorization, resource access, resource discovery, and other challenges are encountered. It is this class of problem that is addressed by Grid technologies. Next, the authors present an extensible and open Grid architecture, in which protocols, services, application programming interfaces, and software development kits are categorized according to their roles in enabling resource sharing. The authors describe requirements that they believe any such mechanisms must satisfy and discuss the importance of defining a compact set of intergrid protocols to enable interoperability among different Grid systems. Finally, the authors discuss how Grid technologies relate to other contemporary technologies, including enterprise integration, application service provider, storage service provider, and peer-to-peer computing. They maintain that Grid concepts and technologies complement and have much to contribute to these other approaches.

The term Gridis used to refer to a software system that provides uniform and location-independent access to geographically and organizationally dispersed, heterogeneous resources that are persistent and supported. These emerging data and computing Grids promise to provide a highly capable and scalable environment for addressing large-scale science problems. The author describes the requirements for science Grids, the resulting services and architecture of NASA’s Information Power Grid and the Department of Energy’s Science Grid, and some of the scaling issues that have come up in their implementation.

The computational Grid is a promising platform for the deployment of large-scale scientific and engineering applications. Parameter sweep applications (PSAs) arise in many fields of science and engineering and are structured as sets of “experiments,” each of which is executed with a distinct set of parameters. Given that structure, PSAs are particularly well suited to the Grid infrastructure and can be deployed on very large scales. However, deployment is not easy to achieve for the domain scientist given the complexity and multiplicity of the Grid software infrastructure, the heterogeneity of the resources, and the dynamic resource availabilities. It is therefore necessary to provide user-level middleware that acts as an intermediate layer between the application and the Grid. That middleware must address all deployment, data movements, and scheduling issues to provide the user with a transparent way of running his or her simulation on the Grid. In this paper, the authors focus on such middleware specifically targeted to a biology application: MCell. After describing the application and its structure, they describe desired usage scenarios on the Grid and identify user requirements, discuss relevant computer science issues, and propose suitable solutions given currently available Grid technologies. The authors then describe a general-purpose user-level middleware project for PSAs—AppLes Parameter Sweep Template—explain how it can be extended to accommodate MCell’s specific requirements, and introduce current work in that direction.

In this paper, the authors investigate G-commerce—computational economies for controlling resource allocation in computational Grid settings. They define hypothetical resource consumers (representing users and Grid-aware applications) and resource producers (representing resource owners who “sell” their resources to the Grid). The authors then measure the efficiency of resource allocation under two different market conditions—commodities markets and auctions—and compare both market strategies in terms of price stability, market equilibrium, consumer efficiency, and producer efficiency. The results indicate that commodities markets are a better choice for controlling Grid resources than previously defined auction strategies.

Future-generation space missions across the solar system to the planets, moons, asteroids, and comets may someday incorporate supercomputers both to expand the range of missions being conducted and to significantly reduce their cost. By performing science computation directly on the spacecraft itself, the amount of data required to be downlinked may be reduced by many orders of magnitude, thus greatly reducing the mass of the resources needed for communication while increasing the quality and quantity of the science achieved. By performing the mission planning in real time directly on the spacecraft, complex and highly responsive missions can be conducted out of range of direct human intervention, and the cost of mission management can be reduced. Through highly replicated computing structures, continued operation can be maintained in the presence of faults by means of graceful degradation. Two classes of systems, reflecting very different strategies of computer system architecture, are actively being pursued by the NASA Jet Propulsion Laboratory (JPL) to take advantage of the opportunity of embedded high performance computing on spacecraft for deep-space missions. Commodity off-the-shelf (COTS) clusters may permit the direct application of commercial computing hardware in loosely coupled ensembles to benefit from the enormous investment of industry in mass-market components. New processor-in-memory (PIM) architectures combine multiple nodes on a single chip of processor-memory pairs exposing the full memory bandwidth. This paper examines the driving issues motivating the use of supercomputing for future deep-space missions and describes two active research projects at NASA JPL that are pursuing both the COTS and PIM strategies for next-generation spaceborne computing.

This paper presents a comparison of three programming models—OpenMP, HPF, and MPI—applied to a diphasic compressible fluid mechanics code. The parallelization analysis is conducted, and the authors also present the experimental results obtained on various platforms: a Compaq Proliant 6000 (4 processors), a Cray T3E-750 (300 processors), an HP Class V (16 processors), a SGI Origin 2000 (32 processors), a cluster of PCs, and a COMPAQ SC 232 (232 processors). These experimental results will be discussed according to the following criteria: efficiency, scalability, maintainability, developing costs, and portability. As a conclusion, the authors present the parallelization strategy recommended for codes comparable to ECOSS.

In this paper, the authors deal with algorithmic issues on heterogeneous platforms. They concentrate on dense linear algebra kernels, such as matrix multiplication or LU decomposition. Block-cyclic distribution techniques used in ScaLAPACK are no longer sufficient to balance the load among processors running at different speeds. The main result of this paper is to provide a static data distribution scheme that leads to an asymptotically perfect load balancing for LU decomposition, thereby providing solid foundations toward the design of a cluster-oriented version of ScaLAPACK.