Fusion reactions between 12C nuclei are among the most important in stellar evolution since they determine the destiny of massive stars (> 8 M). At thermonuclear energies (Ecm=1.5 ± 0.3 MeV), the 12C+12C reactions mainly proceed through 20Ne+α and 23Na+p channels. Since these energies are much lower than the height of the Coulomb barrier (Ecm= 6.1 MeV), the direct measurements of the 12C+12C reactions are very challenging because of the extremely small cross sections involved and the high beam-induced background originating from impurities in the targets. In addition, the 12C+12C reaction forms 24Mg at relatively high excitation energies (above the 12C 13.93 MeV thresholds) where molecular configurations are possible. Theoretical models fail to reproduce such structures and as a result, the extrapolation of high-energy cross section data towards the energy of astrophysical interest remain uncertain by 2-3 orders of magnitude. Further experimental efforts to measure at the lowest accessible energies are therefore in need. However, additionally to the extremely low cross sections and the resonant structure, the measurements at stellar energies of the 12C+12C reactions are troublesome due to natural hydrogen and deuterium contamination in the carbon targets. These target contaminants hamper the measurement of the 12C+12C process in all exit channels given that the 12C+1,2H reactions cross sections are considerably higher than that of the reaction of interest. In consequence, the use of ultra-low H content graphite targets and a study of the target behaviour under beam bombardment are necessary. This work focused on the experimental measurements of the 12C(12C,p)23Na and 12C(12C,α)20Ne reactions using charge particle detection. Although both channels were measured, only the proton channel was analysed and discussed in this thesis due to time constrains. The experiment was performed at the 3 MV pelletron tandem accelerator of the CIRCE (Centre for Isotopic Research on the Cultural and Environmental heritage) laboratory in Caserta, Italy. The experimental approach involved the development of optical calculations for optimal beam transportation (using the software COSY), the use of a four ΔE-Erest detectors system (a variable pressure CF4 ionization chamber used as the ΔE detector and a 300 mm2 Si detector used as the Erest) called GASTLY (GAs Silicon Two-Layer sYstem) and a study of the deuterium (hydrogen does not contribute to the beam-induced background at the detection angles and beam energies used here) contamination in graphite targets. The GASTLY detectors were placed at backward angles (121, 143 and 156o respect to the beam axis) and the 12C+12C reactions were investigated using carbon beams of Ecm=4.30 - 2.52 MeV with intensities of the order of μA. Highly Ordered Pyrolytic Graphite (HOPG) and highly pure (99.8%) natural graphite targets were used for the deuterium contamination study. A thermocamera was used to constantly monitor the target temperature during beam bombardment, allowing the investigation of target's deuterium content as a function of target temperature. Results showed a decrease in target's deuterium content of 53-80% in the target's temperature range of 200-1200 °C, depending on the type of target and detection angle. Furthermore, it was found that surrounding the scattering chamber with a nitrogen atmosphere while measuring low counting rate reactions (such as 12C+12C at low energies), the HOPG target's deuterium content decreases to about half its original value for a target temperature in the range between 800-1100 °C. For the 12C+12C reactions measurements, the HOPG target was used, maintaining high target temperatures. The p0-6 proton groups of the 12C(12C,p)23Na reaction were analysed and their yields, cross sections and astrophysical S-factors were obtained and are presented in this thesis. A comparison with previous data available in the literature is also presented, together with an indication for possible improvements in future investigations.

The Muon Ionisation Cooling Experiment (MICE) has been commissioned to provide the first demonstration of ionisation cooling. MICE will aim to demonstrate that ionisation cooling can be used to reduce of the emittance of a beam of muons to meet the requirements of future particle physics experiments such as the Neutrino Factory, or Muon Collider. As of October 2016, commissioning of Step IV of MICE has been completed which provides an opportune time to make material physics studies on the absorber material. The cooling formula that MICE will use to measure the emittance reduction was reviewed. It is shown that the energy loss term is important when measuring cooling, and an accurate measurement of the energy loss will hence improve the accuracy of the cooling formula. The physics of ionisation cooling is also reviewed. The primary absorber used in the early data taking of MICE Step IV will be a 65mm disk of Lithium Hydride. The energy loss of Lithium Hydride was estimated using the equations of energy loss developed by Bethe. Methods were developed in this thesis to make measurements of the energy loss using data from the MICE trackers, and the timeof- flight data through the cooling channel. The energy loss of muons in monte-carlo simulations measured with the two alternative methods was found to be in agreement, with a measurement by the trackers of 9.02 ± 0.07, and from simulated time-of-flight of 9.32 ± 0.15. The first measurement of energy loss by 200 MeV/c muons was made using time-of-flight data using real muons in the MICE channel of ∆E = 9.23 ± 0.13 MeV, corresponding to a stopping power of Lithium Hydride of dE/dx = 1.42 ± 0.02 MeV g−1 cm2.

The interest in social entrepreneurship is increasing in Europe in general and Sweden in particular. Because social entrepreneurs (SEs) and incubators share a common aim of enhancing development and improving the living conditions of the people, one would think that they tend to work together more closely than in case when incubators work with the conventional entrepreneurs (CEs). Incubator activities can influence the process of the SE. The purpose of this paper was to identify which activities, that they provide, can influence the SE’s growth. We were also interested in examining how those activities differ when working with the CEs. To fulfill this purpose, we developed the research question: What are the roles of incubators and accelerators in the entrepreneurial process of SE and how they differ from the roles played in the entrepreneurial process of CE?  In order to answer the research question and fulfil the objectives of study in hand, we embraced interpretivist approach and qualitative method for data collection and analysis. 6 semi structured interviews are conducted with SEs that have experience in participating in incubator and accelerator programs, as well as two experts from the side of incubator and accelerator programs.  It has been found that education, making contacts and facilitating meetings as well as the office space are the main roles of incubator and accelerator programs, which have as their goal to assist SEs. Moreover, it is found that not all of the roles are equally significant, nor that they are played out through all the phases of the entrepreneurial process. Stages of the entrepreneurial process are revised, and their content is found to be different in comparison with social entrepreneurship. We answer the second part of our research question by carrying out a comparative discussion based on the empirical findings on SEs and existing literature on CEs. We summaries these difference by the means of a table. Finally, we put together and connect roles of incubator and accelerator programs interested in scaling of SEs and the entrepreneurial process, where the model linking the roles through the entrepreneurial phases of SE is proposed. Hence, we conclude that our research question has been answered and research purpose fulfilled.

One of the most promising devices to provide accurate measurement of the longitudinal bunch profile at the tens of femtosecond level is based on electro-optic techniques. In this thesis, a bunch profile monitor, based on electro-optic spectral decoding (EOSD), is currently developed for the CLIC Test Facility 3 at CERN. The monitor is optimised for bunch lengths over 3.5 ps with effective window of 16 ps, and sub-picosecond resolution. The measurement results from the EO monitor are compared with measurements by coherent transition radiation on a streak camera. The measurement on bunch charge dependence is studied. Timing resolution of the bunch profile monitor is studied in both theory and numerical calculation. This thesis summarises a frequency analysis approach of electro-optic effect based on $\chi^{(2)}$ frequency mixing process. From the theory analysed in frequency domain, a non-crossed polarization measurement includes all three of the probe laser background term, the linear term to Coulomb field and the quadratic term to Coulomb field. Three methods are induced based on this frequency analysis result to retrieve Coulomb field value which is emitted from electron beam. The measured 1.3 MV/m field strength agrees with calculation result. An experiment is designed to study the role of incident beam sizes and non-collinear incident beams in EO technique. Due to the phase matching process, the non-collinear angle of the incident beams induces a frequency dependent angular chirp in the beams emitted after the EO crystal. This frequency offset may lead to frequency loss in fibre coupling, and thus lead to bunch length broadening in a measurement for short electron bunch.

abstract: The rapid improvement in computation capability has made deep convolutional neural networks (CNNs) a great success in recent years on many computer vision tasks with significantly improved accuracy. During the inference phase, many applications demand low latency processing of one image with strict power consumption requirement, which reduces the efficiency of GPU and other general-purpose platform, bringing opportunities for specific acceleration hardware, e.g. FPGA, by customizing the digital circuit specific for the deep learning algorithm inference. However, deploying CNNs on portable and embedded systems is still challenging due to large data volume, intensive computation, varying algorithm structures, and frequent memory accesses. This dissertation proposes a complete design methodology and framework to accelerate the inference process of various CNN algorithms on FPGA hardware with high performance, efficiency and flexibility. As convolution contributes most operations in CNNs, the convolution acceleration scheme significantly affects the efficiency and performance of a hardware CNN accelerator. Convolution involves multiply and accumulate (MAC) operations with four levels of loops. Without fully studying the convolution loop optimization before the hardware design phase, the resulting accelerator can hardly exploit the data reuse and manage data movement efficiently. This work overcomes these barriers by quantitatively analyzing and optimizing the design objectives (e.g. memory access) of the CNN accelerator based on multiple design variables. An efficient dataflow and hardware architecture of CNN acceleration are proposed to minimize the data communication while maximizing the resource utilization to achieve high performance. Although great performance and efficiency can be achieved by customizing the FPGA hardware for each CNN model, significant efforts and expertise are required leading to long development time, which makes it difficult to catch up with the rapid development of CNN algorithms. In this work, we present an RTL-level CNN compiler that automatically generates customized FPGA hardware for the inference tasks of various CNNs, in order to enable high-level fast prototyping of CNNs from software to FPGA and still keep the benefits of low-level hardware optimization. First, a general-purpose library of RTL modules is developed to model different operations at each layer. The integration and dataflow of physical modules are predefined in the top-level system template and reconfigured during compilation for a given CNN algorithm. The runtime control of layer-by-layer sequential computation is managed by the proposed execution schedule so that even highly irregular and complex network topology, e.g. GoogLeNet and ResNet, can be compiled. The proposed methodology is demonstrated with various CNN algorithms, e.g. NiN, VGG, GoogLeNet and ResNet, on two different standalone FPGAs achieving state-of-the art performance. Based on the optimized acceleration strategy, there are still a lot of design options, e.g. the degree and dimension of computation parallelism, the size of on-chip buffers, and the external memory bandwidth, which impact the utilization of computation resources and data communication efficiency, and finally affect the performance and energy consumption of the accelerator. The large design space of the accelerator makes it impractical to explore the optimal design choice during the real implementation phase. Therefore, a performance model is proposed in this work to quantitatively estimate the accelerator performance and resource utilization. By this means, the performance bottleneck and design bound can be identified and the optimal design option can be explored early in the design phase. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2018

Neste trabalho apresentamos o projeto do microtron principal do IFUSP. Inicialmente são apresentados os fundamentos teóricos relativos à estabilidade do feixe em um microtron, bem como discussões sobre a ótica do feixe. No trabalho são feitas basicamente 3 tipos de simulações distintas: simulação da dinâmica longitudinal e transversal da etapa síncrona, simulação da dinâmica transversal das linhas de transferência (inserção e extração) e a simulação relativa à otimização do projeto dos eletroímãs principais. As simulações longitudinais mostraram que é possível reduzir a energia de entrada do microtron principal de 4,9 para 2,5 MeV, o que permitira a eliminação do primeiro estágio síncrono (microtron booster). A energia final do microtron principal passa a ser 38 MeV. Foram realizados os projetos de duas linhas de inserção no principal para essas as duas energias (2,5 e 4,9 MeV) e foi feita uma análise comparativa desses dois projetos. O projeto da linha de extração também foi analisado levando em consideração duas energias de extração (5,8 e 38 MeV) para se mostrar a viabilidade dessa etapa. / In this work we present the design of the IFUSP main microtron. Initially, the theoretical foundations for the beam stability in the microton are presented as well as the beam optics is discussed. Three different kinds of simulations were done: Longitudinal and transversal beam dynamics in the synchronous part of the machine, transversal beam dynamics concerning the transfer lines and simulations related to the design of the main magnets. The longitudinal simulations sugested that it is possible to reduce the initial energy of the main microtron from 4.9 to 2.5 MeV, this would permit the elimination of the first synchronous stage (booster microtron). The final energy of the main microtron will be 38 MeV. The design of two insertion lines in the main microtron for the two different energies (2.5 and 4.9 MeV) was done and a comparative analysis between these two designes is presented. The design of the extraction line also takes into account two different extraction energies (5.8 and 38 MeV) in order to show de viability of this part.

Construímos um acelerador de elétrons de baixa energia (opera na faixa de 0,4 - 20 keV) que dispõe dos recursos necessários para ser utilizado como instrumento de pesquisa em áreas relacionadas com os isolantes. Neste trabalho, ele foi empregado para estudar a emissão secundária em polímeros ou, mais especificamente, para mostrar que um novo método de medida de emissão secundária, que designamos \"Método de Medida Dinâmica\", descrito por H. Von Seggern [IEEE Trans. Nucl. Sei. NS-32, p.1503 (1985)] não permite, ao contrário do que se esperava, obter a verdadeira curva de emissão secundária devido à influência exercida pela carga positiva acumulada sobre a emissão. Contudo, no decorrer do trabalho, mostramos que a montagem descrita por ele ainda pode ser utilizada com vantagem para: i) medir precisamente a energia do 2&#176 ponto de cruzamento e ii) para levantar a verdadeira curva de emissão se, ao invés irradiarmos o alvo continuamente, usarmos pulsos de corrente. Além disso, pudemos, analisando o modo como a carga positiva age sobre a emissão nas mais diversas situações. I) estimar a profundidade de escape máxima e média dos secundários; II) mostrar que a carga positiva líquida numa amostra carregada positivamente fica próxima da superfície e III) mostrar que em amostras carregadas negativamente as cargas positivas ficam próximas da superfície e as negativas, em maior número, no volume do material. / We have constructed an accelerator for the generation of low energy e1ectrons (in the 0.4 to 20 keV range). The accelerator is equipped with some devices especially designed for the investigation of the e1ectrical properties of electron-irradiated dielectrics. In this work we have employed it for the study of the secondary electron emission of irradiated polymers. Reference is made to a method proposed bt H. von Seggern [IEEE Trans. Nucl. Sci. NS-32, p.1503 (1985)] which was intended for the determination of the electron emission yield especially between the two cross-over points in a single run, here called the dynamical method. We have been able to prove that, contrary to expectation, this method does not give correct results over the entire emission curve. Rather it gives yield values which are too low by 25% in the region where the emission exhibits a maximum, due to the interaction between the electron emission process and the positive surface charge of the dielectric. However the method needs not to be dismissed entirely. As it is, it can be used advantageously for the precise determination of the energy of the second cross-over point. In addition, with the same set up, the method could be improved by replacing the continuous irradiation of the sample by a pulsed irradiation, leading to results essentially the same as those shown in the literature. Finally, analyzing the process of interaction between the positive charge of the dielectric and the mechanism of electron emission in several situations, we were able: I) to determine the maximum value and the average value of the escape depth of the emitted electrons; II) for a sample with a net positive charge, to show that the positive charge resides very near the surface of incidence; III) for a sample with a net negative charge, to show that the positive charge also resides near the surface while the (prevalent) negative charge resides in the bulk of the material

Neste trabalho apresentamos o projeto, construção e testes de eletroímãs dipolares usados na linha de transporte do feixe do Microtron do IFUSP. Inicialmente são feitos estudos da contribuição do campo de borda dos eletroímãs na curvatura total do feixe e posteriormente esses valores foram usados para a definição de parâmetros de projeto. Foram construídos e caracterizados dois dipolos (30 e 45o) e ainda um eletroímã seletor (+/- 90o) com simetria azimutal. Os resultados obtidos nos testes realizados com os dipolos mostraram um desempenho adequado à operação do acelerador. Também é descrito um método para a simulação da trajetória de elétrons na presença de um campo magnético. / In this work we present the design, construction and testing of dipole magnets used on the beam line of IFUSP-Microtron. Initially we studied the fringe field of the magnets on the total beam bending, these values had been used for the definition of design parameters. It has been constructed two dipoles (30 and 45o) as well as a switch magnet (+/- 90o) with azimuthal symmetry. The tests results of the dipoles had shown an adequate performance to the accelerators operation. A method for the simulation of the electron path in the presence of a magnetic field is also described.

With the upcoming 5G deployment and the exponentially increasing data transmitted over cellular networks, off the shelf hardware won't provide enough performance to cope with the data being transferred over cellular networks. To tackle that problem, hardware accelerators will be of great support thanks to their better performances and lower energy consumption. However, hardware accelerators are not a silver bullet as their very nature prevents them to be as flexible as CPUs. Hardware accelerators integration into Kubernetes and Docker, respectively the most used tools for orchestration and containerization, is still not as flexible as it would need. In this thesis, we developed a framework that allows for a more flexible integration of these accelerators into a Kubernetes cluster using Docker containers making use of an abstraction layer instead of the classic virtualization process. Our results compare the performance of an execution with and without the framework that was developed during this thesis. We found that the framework's overhead depends on the size of the data being processed by the accelerator but does not go over a very low percentage of the total execution time. This framework provides an abstraction for hardware accelerators and thus provides an easy way to integrate hardware accelerated applications into a heterogeneous cluster or even across different clusters with different hardware accelerators types. This framework also moves the hardware specific parts of an accelerated program from the containers to the infrastructure and enables a new kind of service, OpenCL as a service.

Le travail développé dans cette thèse concerne un type particulier de sources de positrons utilisant le rayonnement de canalisation dans un cristal ainsi que d'autres effets cristallins observes le long des axes du cristal ; ces effets produisent un grand nombre de photons qui, à leur tour, génèrent un grand nombre de paires e+e- dans une cible amorphe. Les photons et les paires sont créés dans deux cibles différentes séparées par une certaine distance permettant l'installation d'un aimant pour dévier les particules chargées avant la cible amorphe. Une telle source est appelée source hybride de positrons ; elle a été choisie par le CERN pour le projet CLIC. Ce type de sources présente de réels avantages par rapport aux cibles conventionnelles qui ont une grande emittance ainsi qu'un niveau important de dépôt d'énergie dans la cible. Apres un rappel des phénomènes physiques qui concernent notre étude, des simulations détaillées utilisant d'une part le programme de V .Strakhovenko pour les effets cristallins et d'autre part le code GEANT4 pour la génération des positrons conduisent à une description complète pour les photons et les positrons avec, notamment, les espaces de phase longitudinal et transverse, le spectre en énergie, la distribution temporelle,.. Nous avons particulièrement insiste sur deux points : d'abord sur les dispositifs de capture des positrons -après la cible- qui sont essentiels pour avoir de bons rendements de positrons acceptes et ensuite sur la densité de l'énergie déposée dans la cible qui représente un paramètre important pour la survie des cibles. En ce qui concerne le premier point, trois dispositifs de capture ont été étudiés : le système adiabatique (AMD), le système quart d'onde (QWT) et la lentille de lithium. Pour le deuxième point qui concerne l'énergie déposée et l'échauffement de la cible, on a cherché à optimiser la densité d'énergie déposée en diminuant son maximum (PEDD) ; l'énergie moyenne déposée a aussi été optimisée en utilisant une solution spéciale pour le convertisseur : un convertisseur granulaire forme de petites sphères, comme cela avait été considéré précédemment pour les usines à neutrinos. Des résultats très prometteurs nous ont conduits à envisager la source hybride de positrons avec un convertisseur granulaire comme une solution au difficile problème d'ILC. Cette solution est étudiée moyennant une transformation des impulsions du faisceau avant la cible, comme cela avait été envisage par l'équipe du KEK. Le transport du faisceau de positrons au-delà du solénoïde a été étudié avec la première partie de l'optique quadrupolaire.