GPU accelerated algorithms for computing matrix function vector products with applications to exponential integrators and fractional diffusion

The efficient computation of matrix function vector products has become an important area of research in recent times, driven in particular by two important applications: the numerical solution of fractional partial differential equations and the integration of large systems of ordinary differential equations. In this work we consider a problem that combines these two applications, in the form of a numerical solution algorithm for fractional reaction diffusion equations that after spatial discretisation, is advanced in time using the exponential Euler method. We focus on the efficient implementation of the algorithm on Graphics Processing Units (GPU), as we wish to make use of the increased computational power available with this hardware. We compute the matrix function vector products using the contour integration method in [N. Hale, N. Higham, and L. Trefethen. Computing Aα, log(A), and related matrix functions by contour integrals. SIAM J. Numer. Anal., 46(5):2505–2523, 2008]. Multiple levels of preconditioning are applied to reduce the GPU memory footprint and to further accelerate convergence. We also derive an error bound for the convergence of the contour integral method that allows us to pre-determine the appropriate number of quadrature points. Results are presented that demonstrate the effectiveness of the method for large two-dimensional problems, showing a speedup of more than an order of magnitude compared to a CPU-only implementation.

Contributing to computing services to benchmarking asset management knowledge management

Asset management has broadened from a focus on maintenance management to whole of life cycle asset management requiring a suite of new competencies from asset procurement to management and disposal. Well developed skills and competencies as well as practical experience are a prerequisite to maintain capability, to manage demand as well to plan and set priorities and ensure on-going asset sustainability. This paper has as its focus to establish critical understandings of data, information and knowledge for asset management along with the way in which benchmarking these attributes through computer-aided design may aid a strategic approach to asset management. The paper provides suggestions to improve sharing, integration and creation of asset-related knowledge through the application of codification and personalization approaches.

Distributed computing of all-to-all comparison problems in heterogeneous systems

The requirement of distributed computing of all-to-all comparison (ATAC) problems in heterogeneous systems is increasingly important in various domains. Though Hadoop-based solutions are widely used, they are inefficient for the ATAC pattern, which is fundamentally different from the MapReduce pattern for which Hadoop is designed. They exhibit poor data locality and unbalanced allocation of comparison tasks, particularly in heterogeneous systems. The results in massive data movement at runtime and ineffective utilization of computing resources, affecting the overall computing performance significantly. To address these problems, a scalable and efficient data and task distribution strategy is presented in this paper for processing large-scale ATAC problems in heterogeneous systems. It not only saves storage space but also achieves load balancing and good data locality for all comparison tasks. Experiments of bioinformatics examples show that about 89\% of the ideal performance capacity of the multiple machines have be achieved through using the approach presented in this paper.

A penalty-based genetic algorithm for the migration cost-aware virtual machine placement problem in cloud data centers

In the past few years, the virtual machine (VM) placement problem has been studied intensively and many algorithms for the VM placement problem have been proposed. However, those proposed VM placement algorithms have not been widely used in today's cloud data centers as they do not consider the migration cost from current VM placement to the new optimal VM placement. As a result, the gain from optimizing VM placement may be less than the loss of the migration cost from current VM placement to the new VM placement. To address this issue, this paper presents a penalty-based genetic algorithm (GA) for the VM placement problem that considers the migration cost in addition to the energy-consumption of the new VM placement and the total inter-VM traffic flow in the new VM placement. The GA has been implemented and evaluated by experiments, and the experimental results show that the GA outperforms two well known algorithms for the VM placement problem.

Moving from evaluation to trial: The case of cloud ERP adoption in SMEs

This study proposes that the adoption process of complex-wide systems (e.g. cloud ERP) should be observed as multi-stage actions. Two theoretical lenses were utilised for this study with critical adoption factors identified through the theory of planned behaviour and the progression of each adoption factor observed through Ettlie's (1980) multi-stage adoption model. Together with a survey method, this study has employed data gathered from 162 decision-makers of small and medium-sized enterprises (SMEs). Using both linear and non-linear approaches for the data analysis, the study findings have shown that the level of importance for adoption factors changes across different adoption stages.

The Arctic Summer Cloud Ocean Study (ASCOS): overview and experimental design

The climate in the Arctic is changing faster than anywhere else on earth. Poorly understood feedback processes relating to Arctic clouds and aerosol–cloud interactions contribute to a poor understanding of the present changes in the Arctic climate system, and also to a large spread in projections of future climate in the Arctic. The problem is exacerbated by the paucity of research-quality observations in the central Arctic. Improved formulations in climate models require such observations, which can only come from measurements in situ in this difficult-to-reach region with logistically demanding environmental conditions. The Arctic Summer Cloud Ocean Study (ASCOS) was the most extensive central Arctic Ocean expedition with an atmospheric focus during the International Polar Year (IPY) 2007–2008. ASCOS focused on the study of the formation and life cycle of low-level Arctic clouds. ASCOS departed from Longyearbyen on Svalbard on 2 August and returned on 9 September 2008. In transit into and out of the pack ice, four short research stations were undertaken in the Fram Strait: two in open water and two in the marginal ice zone. After traversing the pack ice northward, an ice camp was set up on 12 August at 87°21' N, 01°29' W and remained in operation through 1 September, drifting with the ice. During this time, extensive measurements were taken of atmospheric gas and particle chemistry and physics, mesoscale and boundary-layer meteorology, marine biology and chemistry, and upper ocean physics. ASCOS provides a unique interdisciplinary data set for development and testing of new hypotheses on cloud processes, their interactions with the sea ice and ocean and associated physical, chemical, and biological processes and interactions. For example, the first-ever quantitative observation of bubbles in Arctic leads, combined with the unique discovery of marine organic material, polymer gels with an origin in the ocean, inside cloud droplets suggests the possibility of primary marine organically derived cloud condensation nuclei in Arctic stratocumulus clouds. Direct observations of surface fluxes of aerosols could, however, not explain observed variability in aerosol concentrations, and the balance between local and remote aerosols sources remains open. Lack of cloud condensation nuclei (CCN) was at times a controlling factor in low-level cloud formation, and hence for the impact of clouds on the surface energy budget. ASCOS provided detailed measurements of the surface energy balance from late summer melt into the initial autumn freeze-up, and documented the effects of clouds and storms on the surface energy balance during this transition. In addition to such process-level studies, the unique, independent ASCOS data set can and is being used for validation of satellite retrievals, operational models, and reanalysis data sets.

Performance analysis of secure unified communications in the VMware-based cloud

Unified communications as a service (UCaaS) can be regarded as a cost-effective model for on-demand delivery of unified communications services in the cloud. However, addressing security concerns has been seen as the biggest challenge to the adoption of IT services in the cloud. This study set up a cloud system via VMware suite to emulate hosting unified communications (UC), the integration of two or more real time communication systems, services in the cloud in a laboratory environment. An Internet Protocol Security (IPSec) gateway was also set up to support network-level security for UCaaS against possible security exposures. This study was aimed at analysis of an implementation of UCaaS over IPSec and evaluation of the latency of encrypted UC traffic while protecting that traffic. Our test results show no latency while IPSec is implemented with a G.711 audio codec. However, the performance of the G.722 audio codec with an IPSec implementation affects the overall performance of the UC server. These results give technical advice and guidance to those involved in security controls in UC security on premises as well as in the cloud.

ELISA: An Estimated Load Information Scheduling Algorithm for Distributed Computing Systems

In this paper, we present a decentralized dynamic load scheduling/balancing algorithm called ELISA (Estimated Load Information Scheduling Algorithm) for general purpose distributed computing systems. ELISA uses estimated state information based upon periodic exchange of exact state information between neighbouring nodes to perform load scheduling. The primary objective of the algorithm is to cut down on the communication and load transfer overheads by minimizing the frequency of status exchange and by restricting the load transfer and status exchange within the buddy set of a processor. It is shown that the resulting algorithm performs almost as well as a perfect information algorithm and is superior to other load balancing schemes based on the random sharing and Ni-Hwang algorithms. A sensitivity analysis to study the effect of various design parameters on the effectiveness of load balancing is also carried out. Finally, the algorithm's performance is tested on large dimensional hypercubes in the presence of time-varying load arrival process and is shown to perform well in comparison to other algorithms. This makes ELISA a viable and implementable load balancing algorithm for use in general purpose distributed computing systems.

Experiments on a plume with off-source heating: Implications for cloud fluid dynamics

We report here on a series of laboratory experiments on plumes, undertaken with the object of simulating the effect of the heat release that occurs in clouds on condensation of water vapor. The experimental technique used for this purpose relies on ohmic heating generated in an electrically conducting plume fluid subjected to a suitable alternating voltage across specified axial stations in the plume flow [Bhat et al., 1989]. The present series of experiments achieves a value of the Richardson number that is toward the lower end of the range that characteristics cumulus clouds. It is found that the buoyancy enhancement due to heating disrupts the eddy structures in the flow and reduces the dilution owing to entrainment of ambient fluid that would otherwise have occurred in the central region of the plume. Heating also reduces the spread rate of the plume, but as it accelerates the flow as well, the overall specific mass flux in the plume does not show a very significant change at the heat input employed in the experiment. However, there is some indication that the entrainment rate (proportional to the streamwise derivative of the mass flux) is slightly higher immediately after heat injection and slightly lower farther downstream. The measurements support a previous proposal for a cloud scenario [Bhat and Narasimha, 1996] and demonstrate how fresh insights into certain aspects of the fluid dynamics of clouds may be derived from the experimental techniques employed here.

Satellite-derived direct radiative effect of aerosols dependent on cloud cover

Aerosols from biomass burning can alter the radiative balance of the Earth by reflecting and absorbing solar radiation(1). Whether aerosols exert a net cooling or a net warming effect will depend on the aerosol type and the albedo of the underlying surface(2). Here, we use a satellite-based approach to quantify the direct, top-of-atmosphere radiative effect of aerosol layers advected over the partly cloudy boundary layer of the southeastern Atlantic Ocean during July-October of 2006 and 2007. We show that the warming effect of aerosols increases with underlying cloud coverage. This relationship is nearly linear, making it possible to define a critical cloud fraction at which the aerosols switch from exerting a net cooling to a net warming effect. For this region and time period, the critical cloud fraction is about 0.4, and is strongly sensitive to the amount of solar radiation the aerosols absorb and the albedo of the underlying clouds. We estimate that the regional-mean warming effect of aerosols is three times higher when large-scale spatial covariation between cloud cover and aerosols is taken into account. These results demonstrate the importance of cloud prediction for the accurate quantification of aerosol direct effects.

Computation of Diffuse Solar Radiation

A technique for computing the spectral and angular (both the zenith and azimuthal) distribution of the solar energy reaching the surface of earth and any other plane in the atmosphere has been developed. Here the computer code LOWTRAN is used for getting the atmospheric transmittances in conjunction with two approximate procedures: one based on the Eddington method and the other on van de Hulst's adding method, for solving the equation of radiative transfer to obtain the diffuse radiation in the cloud-free situation. The aerosol scattering phase functions are approximated by the Hyeney-Greenstein functions. When the equation of radiative transfer is solved using the adding method, the azimuthal and zenith angle dependence of the scattered radiation is evaluated, whereas when the Eddington technique is utilized only the total downward flux of scattered solar radiation is obtained. Results of the diffuse and beam components of solar radiation received on surface of earth compare very well with those computed by other methods such as the more exact calculations using spherical harmonics and when atmospheric conditions corresponding to that prevailing locally in a tropical location (as in India) are used as inputs the computed values agree closely with the measured values.

Architecture Optimization of Aerospace Computing Systems

Simultaneous consideration of both performance and reliability issues is important in the choice of computer architectures for real-time aerospace applications. One of the requirements for such a fault-tolerant computer system is the characteristic of graceful degradation. A shared and replicated resources computing system represents such an architecture. In this paper, a combinatorial model is used for the evaluation of the instruction execution rate of a degradable, replicated resources computing system such as a modular multiprocessor system. Next, a method is presented to evaluate the computation reliability of such a system utilizing a reliability graph model and the instruction execution rate. Finally, this computation reliability measure, which simultaneously describes both performance and reliability, is applied as a constraint in an architecture optimization model for such computing systems. Index Terms-Architecture optimization, computation

Byzantine-resilient distributed computing systems

This paper is aimed at reviewing the notion of Byzantine-resilient distributed computing systems, the relevant protocols and their possible applications as reported in the literature. The three agreement problems, namely, the consensus problem, the interactive consistency problem, and the generals problem have been discussed. Various agreement protocols for the Byzantine generals problem have been summarized in terms of their performance and level of fault-tolerance. The three classes of Byzantine agreement protocols discussed are the deterministic, randomized, and approximate agreement protocols. Finally, application of the Byzantine agreement protocols to clock synchronization is highlighted.

Efficient algorithms for computing Reeb graphs

The Reeb graph tracks topology changes in level sets of a scalar function and finds applications in scientific visualization and geometric modeling. We describe an algorithm that constructs the Reeb graph of a Morse function defined on a 3-manifold. Our algorithm maintains connected components of the two dimensional levels sets as a dynamic graph and constructs the Reeb graph in O(nlogn+nlogg(loglogg)3) time, where n is the number of triangles in the tetrahedral mesh representing the 3-manifold and g is the maximum genus over all level sets of the function. We extend this algorithm to construct Reeb graphs of d-manifolds in O(nlogn(loglogn)3) time, where n is the number of triangles in the simplicial complex that represents the d-manifold. Our result is a significant improvement over the previously known O(n2) algorithm. Finally, we present experimental results of our implementation and demonstrate that our algorithm for 3-manifolds performs efficiently in practice.

Fraunhoffer Diffraction Analysis of Satellite Cloud Imagery

The Fraunhoffer diffraction analysis of cloud-covered satellite imagery has shown that the diffraction pattern follows approximately cosine squared distribution. The overshooting tops of clouds and the shadows cast by them contribute much to the diffraction of light, particularly in the high-frequency range. Indeed, cloud-covered imagery can be distinguished from cloud-free imagery on the basis of rate of decay of the diffracted light power in the high-frequency band.