Sparse systems solving on GPUs with GMRES

Scientific applications very often rely on solving one or more linear systems. When matrices are sparse, iterative methods are preferred to direct ones. Nevertheless, the value of nonzero elements and their distribution (i.e., the sketch of the matrix) greatly influence the efficiency of those methods (in terms of computation time, number of iterations, result precision) or simply prevent the convergence.     

Parallel sparse linear solver with GMRES method using minimization techniques of communications for GPU clusters

In this paper, we aim at exploiting the power computing of a graphics processing unit (GPU) cluster for solving large sparse linear systems. We implement the parallel algorithm of the generalized minimal residual iterative method using the Compute Unified Device Architecture programming language and the MPI parallel environment. The experiments show that a GPU cluster is more efficient than a CPU cluster. In order to optimize the performances, we use a compressed storage format for the sparse vectors and the hypergraph partitioning. These solutions improve the spatial and temporal localization of the shared data between the computing nodes of the GPU cluster.     

Efficient and secure cipher scheme for multimedia contents

Multimedia Tools and Applications - The impact of confidentiality and privacy breaches are more pronounced when dealing with multimedia contents. One of the obvious techniques to counter these...     

Transparent and Robust Silica Coatings with Dual Range Porosity for Enzyme‐Based Optical Biosensing

Hierarchically porous transparent silica coatings combine large specific surface area with enhanced pore accessibility for optical biosensing. This paper describes a versatile approach to fabricate optically transparent silica coatings with multiscale porosity. Thin films (around 1 μm in thickness) of an aqueous suspension of primary silica aggregates form a mesoporous, interconnected matrix, and sacrificial polymer particles template well‐defined, discrete macropores with high structural integrity. The total surface area achieved is around 200 m² g^?1 with mesopore sizes of 20–40 nm and macropores of 250 nm, with a total porosity of 84%. The macro/meso dual range of porosity allows enhanced biocatalyst loadings of l ‐lactate dehydrogenase for detection of lactate. The functionalized films showed a linear response within the range of interest of 1–20 × 10^?3m of lactate. These biosensing coatings therefore strongly enhance sensitivity, speed and reliability of optically based lactate detection as compared to classical thin films with monomodal mesopore structure. Particle‐based simulations and experiments reveal that both the location and connectivity of the macropores control the biosensing performance. The coatings and procedure presented here are versatile, scalable, inexpensive, and are therefore compatible with a wide range of deposition techniques suitable for industrial and health care applications.     

Episodic memory and impairment of an early encoding process in schizophrenia.

Investigations of memory impairment in schizophrenia have frequently revealed a strategic processing deficit at encoding. The authors studied an early encoding process, refreshing (in this case, thinking of a stimulus that has just-previously been presented), and its impact on recognition memory in schizophrenia. Following simultaneous presentation of three words or a single word in the top, middle, or bottom position of the screen, 25 patients with schizophrenia and 25 control participants saw and read a new word (read condition), or a word presented on the previous screen (repeat condition), or saw a dot indicating that they should think of and say the last word to have appeared in that position (refresh condition). Later, on a surprise test, participants were asked to recognize words seen previously and give a Remember, Know, or Guess response according to whether they recognized each on the basis of conscious recollection, familiarity, or guessing. The cognitive operation of refreshing was impaired in schizophrenia: patients were slower on 1-word trials and less accurate on 3-word trials to refresh a word, and their Remember responses did not benefit from refreshing. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Appropriate truncation for photonic crystals

Effective properties of one-dimensional photonic crystals in the resonance domain are investigated. The obtained analytic expressions of effective permittivity and permeability lead to several results. Firstly, in the case of lossless materials, effective permittivity and permeability take, in general, complex values. It is shown that these values are governed by the truncation of the boundary layer. Considering the particular case of a symmetric unit cell, the effective permittivity and permeability become purely real and, by the way, coherent with physics. Finally, in this case with a symmetric unit cell, we show that effective permittivity and permeability can stay nearly constant in a wide range of wavevectors including propagating and evanescent waves.     

Isolation of a Series of Fatty Acid Components of Ongokea gore Seed (Isano) Oil and their Detailed Structural Analysis

The total oil production capacity of isano oil is estimated at about 10,000 tons annually. Previous studies of this oil revealed that it is rich in fatty acids including a conjugated diyne moiety. This makes isano oil an excellent candidate for sustainable applications development. However, only a few of its fatty acids have been isolated and identified so far. In this study, we have reinvestigated this oil by characterizing its physicochemical properties and isolating several of its fatty acids as ethyl esters for their detailed structural analysis and identification. Six ethyl esters of fatty acids constituting isanic oil were isolated by flash column chromatography and semipreparative HPLC. The detailed structural analysis of these fatty acid esters by infrared, high resolution, mass spectroscopy, and nuclear magnetic resonance (1‐D and 2‐D) allowed determining unequivocally their chemical structure. The main fatty acid component of the oil (35.7 %) was identified as isanic acid. Four minor acids were found to possess also two conjugated triple bonds, while the sixth fatty acid does not contain carbon–carbon triple bonds nor double bonds but possessed a cis epoxide function. Results obtained in this study are currently being used to explore potential applications of isano oil.