Energy profile of rollback-recovery strategies in high performance computing

Extreme-scale computing is set to provide the infrastructure for the advances and breakthroughs that will solve some of the hardest problems in science and engineering. However, resilience and energy concerns loom as two of the major challenges for machines at that scale. The number of components that will be assembled in the supercomputers plays a fundamental role in these challenges. First, a large number of parts will substantially increase the failure rate of the system compared to the failure frequency of current machines. Second, those components have to fit within the power envelope of the installation and keep the energy consumption within operational margins. Extreme-scale machines will have to incorporate fault tolerance mechanisms and honor the energy and power restrictions. Therefore, it is essential to understand how fault tolerance and energy consumption interplay. This paper presents a comparative evaluation and analysis of energy consumption of three different rollback-recovery protocols: checkpoint/restart, message logging and parallel recovery. Our experimental evaluation shows parallel recovery has the minimum execution time and energy consumption. Additionally, we present an analytical model that projects parallel recovery can reduce energy consumption more than 37% compared to checkpoint/restart at extreme scale.     

On the observation of inductive high-frequency impedance behaviour during the electrodeposition of Au—Sn alloys

Inductive high-frequency impedance behaviour is often observed in metal electrodeposition systems. This behaviour is typically attributed to equipment limitations or non-idealities in the cell set-up and electrical connections. Such instrumental artefacts would nevertheless be relevant to a frequency range which is expected to be well above that in which inductive behaviour is in fact observed (down to a few tens of Hz). In this paper some results on an acidic Au—Sn electrodeposition system are reported. Electrochemical impedance and potentiostatic transients were measured. These results suggest that the high-frequency inductive behaviour may be related to metal nucleation processes. A correlation is proposed between the pseudo-inductive potentiostatic nucleation transients and the pseudo-inductive behaviour of the impedance spectra.     

Surface properties of fluorinated hybrid coatings

A commercial perfluoropolyether containing alkoxysilane functionalities was employed to prepare organic–inorganic hybrid coatings by using the sol‐gel process in the presence of tetraethoxysilane. Contact angle analysis revealed a strong hydrophobic and oleophobic character of the coatings almost independently from the molecular weight of the starting fluorinated oligomer. Surface tension values were in the range of 14–16 mN/m, suggesting a preferential segregation of fluorinated segments onto the surface of the coating. Atomic force microscopy showed the presence very smooth surfaces permitting to neglect the contribution of the surface roughness to wettability. Friction coefficient values were markedly lower with respect to the value of uncoated glass substrate. © 2006 Wiley Periodicals, Inc. J Appl Polym Sci 102: 1483–1488, 2006     

With a little help from my friends

In this paper, we propose a content-based method the for semi-automatic organization of photo albums based on the analysis of how different users organize their own pictures. The goal is to help the user in dividing his pictures into groups characterized by a similar semantic content. The method is semi-automatic: the user starts to assign labels to the pictures and unlabeled pictures are tagged with proposed labels. The user can accept the recommendation or made a correction. To formulate the suggestions is exploited the knowledge encoded in how other users have partitioned their images. The method is conceptually articulated in two parts. First, we use a suitable feature representation of the images to model the different classes that the users have collected, second, we look for correspondences between the criteria used by the different users. Boosting is used to integrate the information provided by the analysis of multiple users. A quantitative evaluation of the proposed approach is obtained by simulating the amount of user interaction needed to annotate the albums of a set of members of the flickr^® photo-sharing community.     

Manufacturing of polypropylene laminates and related structural reorganization in the crystalline phase

The conditions for obtaining recrystallization in more ordered modifications are described for an uniaxially oriented sample of isotactic polypropylene, annealed with fixed ends. These recrystallization conditions are related to the conditions of a lamination process.     

Modification of porous copolymers network based on acrylonitrile

Summary The preparation of a chelating ion-exchange network based on acrylonitrile was carried out by chemical modification with hydroxylamine. The beads of resin were synthesized by aqueous suspension copolymerization of acrylonitrile (AN), styrene (STY) and divinylbenzene (DVB). The influence of diluent used in the suspension polymerization on the structure of the resulting copolymers was evaluated. The diluents employed were heptane (HEP), toluene (TOL) and anisole (ANI). It was found that the AN incorporation into copolymer structure was dependent on the diluent used. Conversion of nitrile groups into the amidoxime was conducted by treatment with hydroxylamine under alkaline solution. The resins were characterized by apparent density, surface area, average pore diameter, elemental analysis (CHN), FTIR and optical microscopy. Based on the results obtained, it was possible to control the porosity by diluent employed in the synthesis and to modify chemically a resin containing nitrile groups by hydroxylamine reaction. Received: 6 October 2001/Revised version: 2 April 2002/ Accepted: 11 April 2002     

Comparative study of homogeneous and heterogeneous styrene polymerizations with Ni(acac)2/MAO catalytic system

Styrene polymerization was carried out with Ni(acac)₂/MAO and Ni(acac)₂/SiO₂/MAO. The influence of reaction parameters (Al/Ni mole ratio, catalyst concentration, temperature and time polymerization) on styrene polymerization was evaluated. It was observed that both catalytic systems were affected by reaction parameters and that the heterogeneous catalyst presented higher activity than the homogeneous one. Polystyrenes with different molecular weight, stereoregularity and polydispersity were obtained. These results suggest that different active catalyst species could have been present. In addition, two types of methylaluminoxane (MAO) with different molecular weights were also evaluated as cocatalyst. As a result, the catalyst activity and stereospecificity were strongly affected by the MAO type.     

Mechanical and dynamical mechanical properties of chloroprene rubber and cellulose II composites

Vulcanized composites of chloroprene rubber (CR) with cellulose II (Cel II) as a filler were investigated. Cel II, obtained by the coagulation of cellulose xanthate, was incorporated in the rubber by the traditional method. The filler content varied from 0 to 30 phr. For comparison purposes, carbon black (CB)–CR composites were also studied. The CB amount varied from 0 to 45 phr. The mechanical and dynamic mechanical properties were determined, and the CR composite containing 20 phr of Cel II showed the best set of properties. © 2004 Wiley Periodicals, Inc. J Appl Polym Sci 92: 2425–2430, 2004     

Dynamic Parameter Identification Based on Lagrangian Formulation and Servomotor-type Actuators for Industrial Robots

International Journal of Control, Automation and Systems - This article presents the design of a parameter identification model for industrial manipulator robots. This model includes the dynamics...     

A multimodal feature learning approach for sentiment analysis of social network multimedia

In this paper we investigate the use of a multimodal feature learning approach, using neural network based models such as Skip-gram and Denoising Autoencoders, to address sentiment analysis of micro-blogging content, such as Twitter short messages, that are composed by a short text and, possibly, an image. The approach used in this work is motivated by the recent advances in: i) training language models based on neural networks that have proved to be extremely efficient when dealing with web-scale text corpora, and have shown very good performances when dealing with syntactic and semantic word similarities; ii) unsupervised learning, with neural networks, of robust visual features, that are recoverable from partial observations that may be due to occlusions or noisy and heavily modified images. We propose a novel architecture that incorporates these neural networks, testing it on several standard Twitter datasets, and showing that the approach is efficient and obtains good classification results.