Showing papers in "Computing in Science and Engineering in 2010"
TL;DR: The OpenCL standard offers a common API for program execution on systems composed of different types of computational devices such as multicore CPUs, GPUs, or other accelerators as mentioned in this paper, such as accelerators.
Abstract: The OpenCL standard offers a common API for program execution on systems composed of different types of computational devices such as multicore CPUs, GPUs, or other accelerators.
1,227 citations
TL;DR: The HUBzero cyberinfrastructure lets scientific researchers work together online to develop simulation and modeling tools and launch simulation runs on the national Grid infrastructure, without having to download or compile any code.
Abstract: The HUBzero cyberinfrastructure lets scientific researchers work together online to develop simulation and modeling tools. Other researchers can then access the resulting tools using an ordinary Web browser and launch simulation runs on the national Grid infrastructure, without having to download or compile any code.
270 citations
TL;DR: The wide diversity of computer architectures today requires a new approach to software development, and OpenMM is an abstraction layer for molecular mechanics simulations, allowing a single program to run efficiently on a variety of hardware platforms.
Abstract: The wide diversity of computer architectures today requires a new approach to software development. OpenMM is an abstraction layer for molecular mechanics simulations, allowing a single program to run efficiently on a variety of hardware platforms.
206 citations
TL;DR: Roundtable participants identified ways of making computational research details readily available, which is a crucial step in addressing the current credibility crisis.
Abstract: Roundtable participants identified ways of making computational research details readily available, which is a crucial step in addressing the current credibility crisis.
170 citations
TL;DR: An overview of the research project is offered and the first experiments with these test chips running spiking neurons based on Eugene Izhikevich's model are described.
Abstract: SpiNNaker is a massively parallel architecture with more than a million processing cores that can model up to 1 billion spiking neurons in biological real time. Here, we offer an overview of our research project and describe the first experiments with these test chips running spiking neurons based on Eugene Izhikevich's model. Note that we're not targeting artificial neural networks (such as perceptrons or multilayer networks) that were inspired by, but don't model, biologically plausible neural systems.
103 citations
TL;DR: MCALab was developed to demonstrate key MCA concepts and make them available to interested researchers and to demonstrate Reproducible research is essential to give MCA a firm scientific foundation.
Abstract: Morphological component analysis of signals and images has far-reaching applications in science and technology, but some consider it problematic and even intractable. Reproducible research is essential to give MCA a firm scientific foundation. Researchers developed MCALab to demonstrate key MCA concepts and make them available to interested researchers.
82 citations
TL;DR: Identifying low-level bottlenecks-and how to ameliorate them-can save hours of frustration over poor performance in apparently well-written programs.
Abstract: CPUs spend most of their time waiting for data to arrive. Identifying low-level bottlenecks-and how to ameliorate them-can save hours of frustration over poor performance in apparently well-written programs.
69 citations
TL;DR: Unlike other socket-based reconfigurable coprocessors, the Convey HC-1 contains nearly 40 field-programmable gate arrays, scatter-gather memory modules, a high-capacity crossbar switch, and a fully coherent memory system.
Abstract: Unlike other socket-based reconfigurable coprocessors, the Convey HC-1 contains nearly 40 field-programmable gate arrays, scatter-gather memory modules, a high-capacity crossbar switch, and a fully coherent memory system.
57 citations
TL;DR: A brief introduction to electronic structure methods is given and two new tools for accelerating matrix-multiplications where single-precision accuracy is insuffcient are proposed.
Abstract: Graphical processing units are now being used with dramatic effect to accelerate quantum chemistry applications. The authors give a brief introduction to electronic structure methods and describe their efforts to accelerate a correlated quantum chemistry code. They propose and analyze two new tools for accelerating matrix-multiplications where single-precision accuracy is insuffcient.
53 citations
TL;DR: Recent progress in atomic and molecular level modeling and simulation of nanoscale materials and processes, as well as efforts by the US National Science Foundation's Network for Computational Nanotechnology (NCN) to cyber-enable such simulation tools together with instructional materials and research seminars are described.
Abstract: The article describes recent progress in atomic and molecular level modeling and simulation of nanoscale materials and processes, as well as efforts by the US National Science Foundation's Network for Computational Nanotechnology (NCN) to cyber-enable such simulation tools together with instructional materials and research seminars. We believe that making advanced simulation tools widely and easily available to the research and education community will significantly enhance the impact of modeling and simulation on nanoscience and nanotechnology To materialize this vision, NCN established nanoHUB.org, a next-generation Web portal or science gateway that lets users run live, interactive simulations, explore data, and learn-all though a simple Web browser without installing any software or providing compute cycles.
37 citations
TL;DR: The next generation of power-efficient parallel computers requires a rethink of the mechanisms and methodology for building parallel applications, and new execution models are required to span all scales, from desktop to supercomputer.
Abstract: Creating the next generation of power-efficient parallel computers requires a rethink of the mechanisms and methodology for building parallel applications. Energy constraints have pushed us into a regime where parallelism will be ubiquitous rather than limited to highly specialized high-end supercomputers. New execution models are required to span all scales, from desktop to supercomputer.
TL;DR: A new hybrid cosmology simulation code is developed making essential use of hardware acceleration for understanding why the expansion of the Universe is accelerating and what dark matter is made of.
Abstract: The advent of powerful cosmological surveys demands a new generation of high-precision, large-volume, and high dynamic range simulations of structure formation in the Universe. Key aims of these simulations are understanding why the expansion of the Universe is accelerating and what dark matter is made of. The availability of Roadrunner, the world's first petaflop platform, led us to develop a new hybrid cosmology simulation code making essential use of hardware acceleration. We describe the strategies underlying the code and aspects of its implementation.
TL;DR: This model of excitation-contraction coupling in skeletal muscles links cell-level electro-physiological behavior with organ-level biomechanical behavior.
Abstract: Computational models of the human body must be accurate enough to use in hypothesis testing or biological function analysis. This is possible only when such models use physiological information from different scales-such as the cell, tissue, and organ levels. This model of excitation-contraction coupling in skeletal muscles links cell-level electro-physiological behavior with organ-level biomechanical behavior.
TL;DR: This work takes an expensive routine from the CAM atmosphere model, and port it to a GPU using CUDA, and uses the experience gained as a guide in thinking about porting the full application to an accelerator based system.
Abstract: Much success has been achieved using GPUs to accelerate existing applications that are highly data parallel, or that are dominated by small, intense computational kernels. What are the prospects for porting existing large scientific models that do not fit this mold? We take an expensive routine from the CAM atmosphere model, and port it to a GPU using CUDA. We use the experience gained as a guide in thinking about porting the full application to an accelerator based system. We consider the best path forward for getting large scientific models running on accelerator based systems, and identify cases where porting may be feasible, and where a complete redesign may be the best option.
TL;DR: This article looks at how recent developments make it possible to exploitFPGA-based architectures' benefits for large-scale systems targeting compute- and data-intensive applications.
Abstract: FPGA-based architectures are known for their applicability to embedded systems. The article looks at how recent developments make it possible to exploit this technology's benefits for large-scale systems targeting compute- and data-intensive applications.
TL;DR: Researchers have continually developed the Nanoelectronic Modeling (NEMO) toolset to provide insight into nanoscale semiconductor devices that are dominated by quantum mechanical effects, and the resulting insights led to the creation of OMEN, a new simulation engine.
Abstract: Researchers have continually developed the Nanoelectronic Modeling (NEMO) toolset over the past 15 years to provide insight into nanoscale semiconductor devices that are dominated by quantum mechanical effects. The ability to represent realistically large devices on an atomistic basis has been the key element in matching experimental data and guiding experiments. The resulting insights led to the creation of OMEN, a new simulation engine.
TL;DR: This research presents a meta-modelling framework that automates the very labor-intensive and therefore time-heavy and therefore expensive process of manually calculating the optimal solution to optimization problems.
Abstract: Global optimization's active research community continues to pursue increasingly efficient computational strategies for addressing large-scale optimization problems.
TL;DR: Although double precision is usually enough, arbitrary precision increases accuracy and the reproducibility of floating-point computations as mentioned in this paper, and arbitrary precision can be used to increase accuracy and reproduceability of floating point computations.
Abstract: Although double precision is usually enough, arbitrary precision increases accuracy and the reproducibility of floating-point computations.
TL;DR: The various research and applications herein can help foster greater insight and understanding of specific geographic problems and share innovative techniques for addressing these problems in the fast-growing interdisciplinary community.
Abstract: This special issue addresses the challenges in handling geographic problems at a global scale, from earthquakes to climatic change, and from water-resource distribution to mountain classification. The various research and applications herein can help foster greater insight and understanding of specific geographic problems and share innovative techniques for addressing these problems in the fast-growing interdisciplinary community.
TL;DR: Through their simulation-based engineering (SBE) design partnership, Goodyear achieved a substantial competitive advantage in new product development and Sandia National Laboratories was able to solve previously intractable nuclear weapons design problems.
Abstract: Through their simulation-based engineering (SBE) design partnership, Goodyear achieved a substantial competitive advantage in new product development and Sandia National Laboratories was able to solve previously intractable nuclear weapons design problems. However, while other governments invest heavily in SBE for global competitiveness, the US has eliminated technical-transfer funding that was critical to establishing the Goodyear-Sandia partnership.
TL;DR: This paper presents an efficient computational algorithm for studying complex phenomena, invasion percolation, which has found wide applications and is applicable to agriculture, medicine, and engineering.
Abstract: Developing efficient computational algorithms for studying complex phenomena has been a long-standing problem in physics and engineering. One such algorithm, invasion percolation, has found wide applications.
TL;DR: In this paper, service-oriented workflow is used to automate task coordination without direct human control, where each task and legacy scientific code can be presented as a Web service and thereby simplifying workflow development and maintenance.
Abstract: Grid-based simulation usually involves sequences of activities such as resource discovery, job submission, file transfer, analysis and simulation, and data harvesting. We can use service-oriented workflow to automate task coordination without direct human control, where each task and legacy scientific code can be presented as a Web service and thereby simplifying workflow development and maintenance. Because each service is described using Web Service Description Language, workflow manager works only with the WSDL interface rather than specific implementations.
TL;DR: The architectures of a future exascale computing system and a future battery-operated embedded system have in common, but many common power and packaging issues as well.
Abstract: What do the architectures of a future exascale computing system and a future battery-operated embedded system have in common? At first glance, their requirements and challenges seem unrelated. However, discussions and collaboration on the projects revealed not only similar requirements, but many common power and packaging issues as well.
TL;DR: A new exact representation of the error function of real arguments justifies an accurate and simple analytical approximation of this function in physical sciences.
Abstract: A new exact representation of the error function of real arguments justifies an accurate and simple analytical approximation. Two of the most widely used functions in physical sciences are the error function erf(x) and its related complimentary error function erfc(x). These functions occur extensively in problems relating to diffusion, heat conduction, and probability. When the argument x is real, rational approximations for these functions provide a high accuracy. In addition, the Faddeeva function, which is a variant of the error function for the complex argument z = x + iy, is used extensively in nuclear physics and spectroscopy.
TL;DR: This issue of CiSE is based on work presented at a US National Science Foundation workshop, Path to Petascale: Adapting Geo/Chem/Astro Applications for Accelerators and Accelerator Clusters, held at the US National Center for Supercomputing Applications (NCSA) in early 2009.
Abstract: This issue of CiSE is based on work presented at the US National Science Foundation workshop, Path to Petascale: Adapting Geo/Chem/Astro Applications for Accelerators and Accelerator Clusters, held at the US National Center for Supercomputing Applications (NCSA) in early 2009. The workshop was designed to raise awareness about the emergence of accelerator-based high-performance computing (HPC) resources among computational scientists from the geosciences, computational chemistry, and astronomy and astrophysics communities and to help them get started in using these resources.
TL;DR: Numerical results model imaging of the spatially distributed dielectric permittivity function in an environment where antipersonnel landmines are embedded along with stones.
Abstract: How can we differentiate between an underground stone and a landmine? A class of new numerical methods aims to address this question using globally convergent-rather than locally convergent-algorithms for coefficient inverse problems. Numerical results model imaging of the spatially distributed dielectric permittivity function in an environment where antipersonnel landmines are embedded along with stones.
TL;DR: Computational epidemiology researchers are harnessing computing power to crack the complicated mystery of how diseases spread.
Abstract: Born from a desire to predict the future, epidemiology has largely been limited to studying the past. Now, computational epidemiology researchers are harnessing computing power to crack the complicated mystery of how diseases spread.
TL;DR: Quantum mechanical electronic structure calculations are playing an ever-expanding role in advancing nanotechnology as well as in advancing the understanding and design of new functional materials for photovoltaic applications.
Abstract: Quantum mechanical electronic structure calculations are playing an ever-expanding role in advancing nanotechnology as well as in advancing our understanding and design of new functional materials Recent research utilizing quantum mechanical electronic structure calculations is helping to improve upon our understanding of existing nanomaterials-and predict new nanomaterials-for photovoltaic applications
TL;DR: The Cell Broadband Engine as mentioned in this paper is a heterogeneous chip multiprocessor that combines a PowerPC processor core with eight single-instruction multiple-data accelerator cores and delivers high performance on many computationally intensive codes.
Abstract: The Cell Broadband Engine is a heterogeneous chip multiprocessor that combines a PowerPC processor core with eight single-instruction multiple-data accelerator cores and delivers high performance on many computationally intensive codes.
TL;DR: This research presentsField-programmable gate arrays, which are more flexible than application-specific integrated circuits, have emerged as a low-power alternative to CPUs.
Abstract: Field-programmable gate arrays, which are more flexible than application-specific integrated circuits, have emerged as a low-power alternative to CPUs.