HpMC: An Energy- Aware Management System for Multi-Level Memory Architectures

DRAM technology faces density and power challenges to increase capacity because of limitations of physical cell design. To overcome these limitations, system designers are exploring alternative solutions that combine DRAM and emerging NVRAM technologies. Previous work on heterogeneous memories focuses, mainly, on two system designs: PCache, a hierarchical, inclusive memory system, and HRank, a flat, non-inclusive memory system. We demonstrate that neither of these designs can universally achieve high performance and energy efficiency across a suite of HPC workloads. In this work, we investigate the impact of a number of multilevel memory designs on the performance, power, and energy consumption of applications. To achieve this goal and overcome the limited number of available tools to study heterogeneous memories, we created HMsim, an infrastructure that enables n-level, heterogeneous memory studies by leveraging existing memory simulators. We, then, propose HpMC, a new memory controller design that combines the best aspects of existing management policies to improve performance and energy. Our energy-aware memory management system dynamically switches between PCache and HRank based on the temporal locality of applications. Our results show that HpMC reduces energy consumption from 13% to 45% compared to PCache and HRank, while providing the same bandwidth and higher capacity than a conventional DRAM system.

Laser-driven x-ray and neutron source development for industrial applications of plasma accelerators

Pulsed beams of energetic x-rays and neutrons from intense laser interactions with solid foils are promising for applications where bright, small emission area sources, capable of multi-modal delivery are ideal. Possible end users of laser-driven multi-modal sources are those requiring advanced non-destructive inspection techniques in industry sectors of high value commerce such as aerospace, nuclear and advanced manufacturing. We report on experimental work that demonstrates multi-modal operation of high power laser-solid interactions for neutron and x-ray beam generation. Measurements and Monte Carlo radiation transport simulations show that neutron yield is increased by a factor ∼2 when a 1 mm copper foil is placed behind a 2 mm lithium foil, compared to using a 2 cm block of lithium only. We explore x-ray generation with a 10 picosecond drive pulse in order to tailor the spectral content for radiography with medium density alloy metals. The impact of using >1 ps pulse duration on laser-accelerated electron beam generation and transport is discussed alongside the optimisation of subsequent bremsstrahlung emission in thin, high atomic number target foils. X-ray spectra are deconvolved from spectrometer measurements and simulation data generated using the GEANT4 Monte Carlo code. We also demonstrate the unique capability of laser-driven x-rays in being able to deliver single pulse high spatial resolution projection imaging of thick metallic objects. Active detector radiographic imaging of industrially relevant sample objects with a 10 ps drive pulse is presented for the first time, demonstrating that features of 200 μm size are resolved when projected at high magnification.

Rocking behaviour of multi-block columns subjected to pulse-type ground motion accelerations

Ancient columns, made with a variety of materials such as marble, granite, stone or masonry are an important part of the
European cultural heritage. In particular columns of ancient temples in Greece and Sicily which support only the architrave are
characterized by small axial load values. This feature together with the slenderness typical of these structural members clearly
highlights as the evaluation of the rocking behaviour is a key aspect of their safety assessment and maintenance. It has to be noted
that the rocking response of rectangular cross-sectional columns modelled as monolithic rigid elements, has been widely investigated
since the first theoretical study carried out by Housner (1963). However, the assumption of monolithic member, although being
widely used and accepted for practical engineering applications, is not valid for more complex systems such as multi-block columns
made of stacked stone blocks, with or without mortar beds. In these cases, in fact, a correct analysis of the system should consider
rocking and sliding phenomena between the individual blocks of the structure. Due to the high non-linearity of the problem, the
evaluation of the dynamic behaviour of multi-block columns has been mostly studied in the literature using a numerical approach
such as the Discrete Element Method (DEM). This paper presents an introductory study about a proposed analytical-numerical
approach for analysing the rocking behaviour of multi-block columns subjected to a sine-pulse type ground motion. Based on the
approach proposed by Spanos (2001) for a system made of two rigid blocks, the Eulero-Lagrange method to obtain the motion
equations of the system is discussed and numerical applications are performed with case studies reported in the literature and with a
real acceleration record. The rocking response of single block and multi-block columns is compared and considerations are made
about the overturning conditions and on the effect of forcing function’s frequency.
.

Selective Multi-Carrier Index Keying OFDM: Error Propagation Rate with Moment Generating Function

We propose a new selective multi-carrier index keying in orthogonal frequency division multiplexing (OFDM) systems that opportunistically modulate both a small subset of sub-carriers and their indices. Particularly, we investigate the performance enhancement in two cases of error propagation sensitive and compromised deviceto-device (D2D) communications. For the performance evaluation, we focus on analyzing the error propagation probability (EPP) introducing the exact and upper bound expressions on the detection error probability, in the presence of both imperfect and perfect detection of active multi-carrier indices. The average EPP results in closedform are generalized for various fading distribution using the moment generating function, and our numerical results clearly show that the proposed approach is desirable for reliable and energy-efficient D2D applications.

Cooperative communications in multi-rate IEEE 802.11 networks

Esta tese apresenta um estudo sobre alguns dos protocolos de cooperação MAC para redes sem fios utilizando o sistema IEEE 802.11 multi-débito. É proposto um novo modelo de arquitetura para a categorização e análise da cooperação em redes sem fios, tendo este modelo sido aplicado a protocolos cooperativos existentes para camada MAC. É investigado como as características do meio físico, assim como os requisitos de níveis superiores podem ser aplicados ao processo de cooperação, com vista a melhorar as características de funcionamento da rede de comunicações. Para este propósito são exploradas as métricas mais relevantes para o processo de cooperação. São igualmente estudados os limites impostos pelos protocolos da camada MAC e as limitações práticas impostas por protocolos da família de normas que compõem o IEEE 802.11. Neste trabalho foi criada uma métrica multicamada, que permite considerar os requisitos aplicacionais de performance e o tipo de tráfego, assim como a mobilidade dos dispositivos, no funcionamento dos mecanismos de cooperação. Como forma de validação, e para corretamente avaliar o impacto da métrica, um novo protocolo de cooperação foi desenvolvido e implementado. O seu funcionamento é descrito de forma analítica assim como validado através de a um ambiente de simulação. Os resultados obtidos mostram que a utilização de uma métrica multicamada é uma técnica robusta, fornecendo melhorias consistentes no contexto de redes IEEE 802.11. São igualmente demonstradas várias outras características de funcionamento com impacto para as comunicações. Estes dados fornecem uma visão real e encorajadora para a realização de mais pesquisas para a melhoria da performance dos protocolos cooperativos, assim como a sua utilização num variado número de aplicações futuras. No final do documento são apresentados alguns desafios para a continuação da investigação deste tópico.

Photodynamic inactivation of bacteria by cationic porphyrins : their cellular targets and potential environmental applications

Photodynamic inactivation (PDI) is defined as the process of cell destruction by oxidative stress resulting from the interaction between light and a photosensitizer (PS), in the presence of molecular oxygen. PDI of bacteria has been extensively studied in recent years, proving to be a promising alternative to conventional antimicrobial agents for the treatment of superficial and localized infections. Moreover, the applicability of PDI goes far beyond the clinical field, as its potential use in water disinfection, using PS immobilized on solid supports, is currently under study. The aim of the first part of this work was to study the oxidative modifications in phospholipids, nucleic acids and proteins of Escherichia coli and Staphylococcus warneri, subjected to photodynamic treatment with cationic porphyrins. The aims of the second part of the work were to study the efficiency of PDI in aquaculture water and the influence of different physicalchemical parameters in this process, using the Gram-negative bioluminescent bacterium Vibrio fischeri, and to evaluate the possibility of recycling cationic PS immobilized on magnetic nanoparticles. To study the oxidative changes in membrane phospholipids, a lipidomic approach has been used, combining chromatographic techniques and mass spectrometry. The FOX2 assay was used to determine the concentration of lipid hydroperoxides generated after treatment. The oxidative modifications in the proteins were analyzed by one-dimensional polyacrylamide gel electrophoresis (SDS-PAGE). Changes in the intracellular nucleic acids were analyzed by agarose gel electrophoresis and the concentration of doublestranded DNA was determined by fluorimetry. The oxidative changes of bacterial PDI at the molecular level were analyzed by infrared spectroscopy. In laboratory tests, bacteria (108 CFU mL-1) were irradiated with white light (4.0 mW cm-2) after incubation with the PS (Tri-Py+-Me-PF or Tetra-Py+-Me) at concentrations of 0.5 and 5.0 μM for S. warneri and E. coli, respectively. Bacteria were irradiated with different light doses (up to 9.6 J cm-2 for S. warneri and up to 64.8 J cm-2 for E. coli) and the changes were evaluated throughout the irradiation time. In the study of phospholipids, only the porphyrin Tri-Py+-Me-PF and a light dose of 64.8 J cm-2 were tested. The efficiency of PDI in aquaculture has been evaluated in two different conditions: in buffer solution, varying temperature, pH, salinity and oxygen concentration, and in aquaculture water samples, to reproduce the conditions of PDI in situ. The kinetics of the process was determined in realtime during the experiments by measuring the bioluminescence of V. fischeri (107 CFU mL-1, corresponding to a level of bioluminescence of 105 relative light units). A concentration of 5.0 μM of Tri-Py+-Me-PF was used in the experiments with buffer solution, and 10 to 50 μM in the experiments with aquaculture water. Artificial white light (4.0 mW cm-2) and solar irradiation (40 mW cm-2) were used as light sources.

Cortical object segregation and categorization by multi-scale line and edge coding

In this paper we present an improved scheme for line and edge detection in cortical area V1, based on responses of simple and complex cells, truly multi-scale with no free parameters. We illustrate the multi-scale representation for visual reconstruction, and show how object segregation can be achieved with coarse-to-finescale groupings. A two-level object categorization scenario is tested in which pre-categorization is based on coarse scales only, and final categorization on coarse plus fine scales. Processing schemes are discussed in the framework of a complete cortical architecture.

Building the what and where systems: multi-scale lines, edges and keypoints

Computer vision for realtime applications requires tremendous computational power because all images must be processed from the first to the last pixel. Ac tive vision by probing specific objects on the basis of already acquired context may lead to a significant reduction of processing. This idea is based on a few concepts from our visual cortex (Rensink, Visual Cogn. 7, 17-42, 2000): (1) our physical surround can be seen as memory, i.e. there is no need to construct detailed and complete maps, (2) the bandwidth of the what and where systems is limited, i.e. only one object can be probed at any time, and (3) bottom-up, low-level feature extraction is complemented by top-down hypothesis testing, i.e. there is a rapid convergence of activities in dendritic/axonal connections.

Fast and accurate multi-scale keypoints based on end-stopped cells

Increasingly more applications in computer vision employ interest points. Algorithms like SIFT and SURF are all based on partial derivatives of images smoothed with Gaussian filter kemels. These algorithrns are fast and therefore very popular.

Optimization of molecular dynamics simulation code and applications to biomolecular systems

Tese de doutoramento, Bioquimica, Faculdade de Ciências e Tecnologia, Universidade do Algarve, 2015

Etude de la structuration laser femtoseconde multi-échelle de verres d’oxydes dopés à l’argent

La structuration laser femtoseconde de verres d’oxydes est aujourd’hui un domaine de recherche en pleine expansion. L’interaction laser-matière est de plus en plus utilisée pour sa facilité de mise en œuvre et les nombreuses applications qui découlent de la fabrication des composants photoniques, déjà utilisés dans l’industrie des hautes technologies. En effet, un faisceau d’impulsions ultracourtes focalisé dans un matériau transparent atteint une intensité suffisante pour modifier la matière en trois dimensions sur des échelles micro et nanométriques. Cependant, l’interaction laser-matière à ces régimes d’intensité n’est pas encore complètement maîtrisée, et les matériaux employés ne sont pas entièrement adaptés aux nouvelles applications photoniques. Par ce travail de thèse, nous nous efforçons donc d’apporter des réponses à ces interrogations. Le mémoire est articulé autour de deux grands volets. Le premier aborde la question de l’interaction de surface de verres avec de telles impulsions lumineuses qui mènent à l’auto-organisation périodique de la matière structurée. L’influence du dopage en ions photosensibles et des paramètres d’irradiation est étudiée afin d’appuyer et de conforter le modèle d’incubation pour la formation de nanoréseaux de surface. À travers une approche innovante, nous avons réussi à apporter un contrôle de ces structures nanométriques périodiques pour de futures applications. Le second volet traite de cristallisation localisée en volume induite en grande partie par l’interaction laser-matière. Plusieurs matrices vitreuses, avec différents dopages en sel d’argent, ont été étudiées pour comprendre les mécanismes de précipitation de nanoparticules d’argent. Ce travail démontre le lien entre la physicochimie de la matrice vitreuse et le caractère hors équilibre thermodynamique de l’interaction qui influence les conditions de nucléation et de croissance de ces nano-objets. Tous ces résultats sont confrontés à des modélisations de la réponse optique du plasmon de surface des nanoparticules métalliques. Les nombreuses perspectives de ce travail ouvrent sur de nouvelles approches quant à la caractérisation, aux applications et à la compréhension de l’interaction laser femtoseconde pour l’inscription directe de briques photoniques dans des matrices vitreuses.

Radiological protection and nuclear engineering studies in multi-MW target systems

Tese de doutoramento, Física, Universidade de Lisboa, Faculdade de Ciências, 2014

In-situ development of self-defensive antibacterial biomaterials: phenol-g-keratin-EC based bio-composites with characteristics for biomedical applications

Recently, the development of highly inspired biomaterials with multi-functional characteristics has gained considerable attention, especially in biomedical, and other health-related areas of the modern world. It is well-known that the lack of antibacterial potential has significantly limited biomaterials for many challenging applications such as infection free wound healing and/or tissue engineering etc. In this perspective, herein, a series of novel bio-composites with natural phenols as functional entities and keratin-EC as a base material were synthesised by laccase-assisted grafting. Subsequently, the resulting composites were removed from their respective casting surfaces, critically evaluated for their antibacterial and biocompatibility features and information is also given on their soil burial degradation profile. In-situ synthesised phenol-g-keratin-EC bio-composites possess strong anti-bacterial activity against Gram-positive and Gram-negative bacterial strains i.e., B. subtilis NCTC 3610, P. aeruginosa NCTC 10662, E. coli NTCT 10418 and S. aureus NCTC 6571. More specifically, 10HBA-g-keratin-EC and 20T-g-keratin-EC composites were 100% resistant to colonisation against all of the aforementioned bacterial strains, whereas, 15CA-g-keratin-EC and 15GA-g-keratin-EC showed almost negligible colonisation up to a variable extent. Moreover, at various phenolic concentrations used, the newly synthesised composites remained cytocompatible with human keratinocyte-like HaCaT, as an obvious cell ingrowth tendency was observed and indicated by the neutral red dye uptake assay. From the degradation point of view, an increase in the degradation rate was recorded during their soil burial analyses. Our investigations could encourage greater utilisation of natural materials to develop bio-composites with novel and sophisticated characteristics for potential applications.

Multi-agent based electricity market simulator with VPP: Conceptual and implementation issues

This paper presents a new architecture for the MASCEM, a multi-agent electricity market simulator. This is implemented in a Prolog which is integrated in the JAVA program by using the LPA Win-Prolog Intelligence Server (IS) provides a DLL interface between Win-Prolog and other applications. This paper mainly focus on the MASCEM ability to provide the means to model and simulate Virtual Power Producers (VPP). VPPs are represented as a coalition of agents, with specific characteristics and goals. VPPs can reinforce the importance of these generation technologies making them valuable in electricity markets.