Use of shared memory in the context of embedded multi-core processors: exploration of the technology and its limits

Modern embedded systems embrace many-core shared-memory designs. Due to constrained power and area budgets, most of them feature software-managed scratchpad memories instead of data caches to increase the data locality. It is therefore programmers’ responsibility to explicitly manage the memory transfers, and this make programming these platform cumbersome. Moreover, complex modern applications must be adequately parallelized before they can the parallel potential of the platform into actual performance. To support this, programming languages were proposed, which work at a high level of abstraction, and rely on a runtime whose cost hinders performance, especially in embedded systems, where resources and power budget are constrained. This dissertation explores the applicability of the shared-memory paradigm on modern many-core systems, focusing on the ease-of-programming. It focuses on OpenMP, the de-facto standard for shared memory programming. In a first part, the cost of algorithms for synchronization and data partitioning are analyzed, and they are adapted to modern embedded many-cores. Then, the original design of an OpenMP runtime library is presented, which supports complex forms of parallelism such as multi-level and irregular parallelism. In the second part of the thesis, the focus is on heterogeneous systems, where hardware accelerators are coupled to (many-)cores to implement key functional kernels with orders-of-magnitude of speedup and energy efficiency compared to the “pure software” version. However, three main issues rise, namely i) platform design complexity, ii) architectural scalability and iii) programmability. To tackle them, a template for a generic hardware processing unit (HWPU) is proposed, which share the memory banks with cores, and the template for a scalable architecture is shown, which integrates them through the shared-memory system. Then, a full software stack and toolchain are developed to support platform design and to let programmers exploiting the accelerators of the platform. The OpenMP frontend is extended to interact with it.

Mitochondrial respiratory supercomplex association limits production of reactive oxygen species from Complex I

Evidence accumulated in the last ten years has demonstrated that a large proportion of the mitochondrial respiratory chain complexes in a variety of organisms is arranged in supramolecular assemblies called supercomplexes or respirasomes. Besides conferring a kinetic advantage (substrate channeling) and being required for the assembly and stability of Complex I, indirect considerations support the view that supercomplexes may also prevent excessive formation of reactive oxygen species (ROS) from the respiratory chain. Following this line of thought we have decided to directly investigate ROS production by Complex I under conditions in which the complex is arranged as a component of the supercomplex I1III2 or it is dissociated as an individual enzyme. The study has been addressed both in bovine heart mitochondrial membranes and in reconstituted proteoliposomes composed of complexes I and III in which the supramolecular organization of the respiratory assemblies is impaired by: (i) treatment either of bovine heart mitochondria or liposome-reconstituted supercomplex I-III with dodecyl maltoside; (ii) reconstitution of Complexes I and III at high phospholipids to protein ratio. The results of this investigation provide experimental evidence that the production of ROS is strongly increased in either model; supporting the view that disruption or prevention of the association between Complex I and Complex III by different means enhances the generation of superoxide from Complex I . This is the first demonstration that dissociation of the supercomplex I1III2 in the mitochondrial membrane is a cause of oxidative stress from Complex I. Previous work in our laboratory demonstrated that lipid peroxidation can dissociate the supramolecular assemblies; thus, here we confirm that preliminary conclusion that primary causes of oxidative stress may perpetuate reactive oxygen species (ROS) generation by a vicious circle involving supercomplex dissociation as a major determinant.

Stacked searches for high-energy neutrinos from blazars with IceCube

In the year 2013, the detection of a diffuse astrophysical neutrino flux with the IceCube neutrino telescope – constructed at the geographic South Pole – was announced by the IceCube collaboration. However, the origin of these neutrinos is still unknown as no sources have been identified to this day. Promising neutrino source candidates are blazars, which are a subclass of active galactic nuclei with radio jets pointing towards the Earth. In this thesis, the neutrino flux from blazars is tested with a maximum likelihood stacking approach, analyzing the combined emission from uniform groups of objects. The stacking enhances the sensitivity w.r.t. the still unsuccessful single source searches. The analysis utilizes four years of IceCube data including one year from the completed detector. As all results presented in this work are compatible with background, upper limits on the neutrino flux are given. It is shown that, under certain conditions, some hadronic blazar models can be challenged or even rejected. Moreover, the sensitivity of this analysis – and any other future IceCube point source search – was enhanced by the development of a new angular reconstruction method. It is based on a detailed simulation of the photon propagation in the Antarctic ice. The median resolution for muon tracks, induced by high-energy neutrinos, is improved for all neutrino energies above IceCube’s lower threshold at 0.1TeV. By reprocessing the detector data and simulation from the year 2010, it is shown that the new method improves IceCube’s discovery potential by 20% to 30% depending on the declination.

Measurement of the ZZ production cross section and limits to the anomalous triple gauge couplings with forward electrons with the ATLAS detector

Measurements of the self coupling between bosons are important to test the electroweak sector of the Standard Model (SM). The production of pairs of Z bosons through the s-channel is forbidden in the SM. The presence of physics, beyond the SM, could lead to a deviation of the expected production cross section of pairs of Z bosons due to the so called anomalous Triple Gauge Couplings (aTGC). Proton-proton data collisions at the Large Hadron Collider (LHC) recorded by the ATLAS detector at a center of mass energy of 8 TeV were analyzed corresponding to an integrated luminosity of 20.3 fb-1. Pairs of Z bosons decaying into two electron-positron pairs are searched for in the data sample. The effect of the inclusion of detector regions corresponding to high values of the pseudorapidity was studied to enlarge the phase space available for the measurement of the ZZ production. The number of ZZ candidates was determined and the ZZ production cross section was measured to be: rn7.3±1.0(Stat.)±0.4(Sys.)±0.2(lumi.)pb, which is consistent with the SM expectation value of 7.2±0.3pb. Limits on the aTGCs were derived using the observed yield, which are twice as stringent as previous limits obtained by ATLAS at a center of mass energy of 7 TeV.

Efficiency limits for silicon/perovskite tandem solar cells: a theoretical model

The primary goal of this work is related to the extension of an analytic electro-optical model. It will be used to describe single-junction crystalline silicon solar cells and a silicon/perovskite tandem solar cell in the presence of light-trapping in order to calculate efficiency limits for such a device. In particular, our tandem system is composed by crystalline silicon and a perovskite structure material: metilammoniumleadtriiodide (MALI). Perovskite are among the most convenient materials for photovoltaics thanks to their reduced cost and increasing efficiencies. Solar cell efficiencies of devices using these materials increased from 3.8% in 2009 to a certified 20.1% in 2014 making this the fastest-advancing solar technology to date. Moreover, texturization increases the amount of light which can be absorbed through an active layer. Using Green’s formalism it is possible to calculate the photogeneration rate of a single-layer structure with Lambertian light trapping analytically. In this work we go further: we study the optical coupling between the two cells in our tandem system in order to calculate the photogeneration rate of the whole structure. We also model the electronic part of such a device by considering the perovskite top cell as an ideal diode and solving the drift-diffusion equation with appropriate boundary conditions for the silicon bottom cell. We have a four terminal structure, so our tandem system is totally unconstrained. Then we calculate the efficiency limits of our tandem including several recombination mechanisms such as Auger, SRH and surface recombination. We focus also on the dependence of the results on the band gap of the perovskite and we calculare an optimal band gap to optimize the tandem efficiency. The whole work has been continuously supported by a numerical validation of out analytic model against Silvaco ATLAS which solves drift-diffusion equations using a finite elements method. Our goal is to develop a simpler and cheaper, but accurate model to study such devices.

Selective retina therapy (SRT) in patients with geographic atrophy due to age-related macular degeneration

For geographic atrophy (GA) due to age-related macular degeneration (AMD) there is so far no approved treatment option. Usually, increased autofluorescence (AF) levels of different patterns adjacent to the atrophic area indicate lipofuscin-laden retinal pigment epithelium (RPE) cells at a high risk for apoptosis. Herein, SRT was used to selectively treat these cells to stimulate RPE proliferation, in order to reduce or ideally stop further growth of the atrophic area.

[Imaging diagostics of geographic atrophy]

The development of imaging technologies has contributed to the understanding of the genesis and pathophysiological mechanisms of geographic atrophy (GA) secondary to age-related macular degeneration (AMD). Fundus autofluorescence (FAF) imaging allows accurate discrimination of the boundaries of atrophic patches. Furthermore, predictive markers for disease progression can be identified. Non-invasive FAF imaging now represents the gold standard for evaluating progressive enlargement of atrophic areas. By means of high resolution optical coherence tomography (OCT) microstructural retinal changes in GA can be identified. Anatomical endpoints are now being used in interventional GA trials and represent meaningful outcome parameters as surrogate markers in an overall slowly progressive disease which may not affect the fovea until later stages of the disease.

Insights on the role of diabetes and geographic variation in patients with critical limb ischaemia

Patients with critical limb ischaemia (CLI) unsuitable for revascularisation have a high rate of amputation and mortality (30% and 25% at 1 year, respectively). Localised gene therapy using plasmid DNA encoding acidic fibroblast growth factor (NV1FGF, riferminogene pecaplasmid) has showed an increased amputation-free survival in a phase II trial. This article provides the rationale, design and baseline characteristics of CLI patients enrolled in the pivotal phase III trial (EFC6145/TAMARIS).

Blue Light versus Green Light Autofluorescence: Lesion Size of Areas with Geographic Atrophy

Blue-light fundus autofluorescence (FAF) imaging is currently widely used for assessing dry age-related macular degeneration (ARMD). However, at this wavelength, the fovea appears as circular zone of marked hypofluorescence, due to the absorption of macular pigment (MP). This dark spot could be misinterpreted as an atrophic area and could lead to difficulties in identifying small, central changes. The purpose of the study was to analyze differences in image quality, FAF patterns, and lesion size, when using conventional blue-light (Λ(1) = 488 nm) and green-light (Λ(2) = 514 nm) FAF.

Progression of age-related geographic atrophy: role of the fellow eye

PURPOSE. To evaluate the role of fellow eye status in determining progression of geographic atrophy (GA) in patients with age-related macular degeneration (AMD). METHODS. A total of 300 eyes with GA of 193 patients from the prospective, longitudinal, natural history FAM Study were classified into three groups according to the AMD manifestation in the fellow eye at baseline examination: (1) bilateral GA, (2) early/intermediate AMD, and (3) exudative AMD. GA areas were quantified based on fundus autofluorescence images using a semiautomated image-processing method, and progression rates (PR) were estimated using two-level, linear, mixed-effects models. RESULTS. Crude GA-PR in the bilateral GA group (mean, 1.64 mm(2)/y; 95% CI, 1.478-1.803) was significantly higher than in the fellow eye early/intermediate group (0.74 mm(2)/y, 0.146-1.342). Although there was a significant difference in baseline GA size (P = 0.0013, t-test), and there was a significant increase in GA-PR by 0.11 mm(2)/y (0.05-0.17) per 1 disc area (DA; 2.54 mm(2)), an additional mean change of -0.79 (-1.43 to -0.15) was given to the PR beside the effect of baseline GA size. However, this difference was only significant when GA size was ?1 DA at baseline with a GA-PR of 1.70 mm(2)/y (1.54-1.85) in the bilateral and 0.95 mm(2)/y (0.37-1.54) in the early/intermediate group. There was no significant difference in PR compared with that in the fellow eye exudative group. CONCLUSIONS. The results indicate that the AMD manifestation of the fellow eye at baseline serves as an indicator for disease progression in eyes with GA ? 1 DA. Predictive characteristics not only contribute to the understanding of pathophysiological mechanisms, but also are useful for the design of future interventional trials in GA patients.

Geographic information system-based decision support for soil conservation planning in Tajikistan

Soil erosion on sloping agricultural land poses a serious problem for the environment, as well as for production. In areas with highly erodible soils, such as those in loess zones, application of soil and water conservation measures is crucial to sustain agricultural yields and to prevent or reduce land degradation. The present study, carried out in Faizabad, Tajikistan, was designed to evaluate the potential of local conservation measures on cropland using a spatial modelling approach to provide decision-making support for the planning of spatially explicit sustainable land use. A sampling design to support comparative analysis between well-conserved units and other field units was established in order to estimate factors that determine water erosion, according to the Revised Universal Soil Loss Equation (RUSLE). Such factor-based approaches allow ready application using a geographic information system (GIS) and facilitate straightforward scenario modelling in areas with limited data resources. The study showed first that assessment of erosion and conservation in an area with inhomogeneous vegetation cover requires the integration of plot-based cover. Plot-based vegetation cover can be effectively derived from high-resolution satellite imagery, providing a useful basis for plot-wise conservation planning. Furthermore, thorough field assessments showed that 25.7% of current total cropland is covered by conservation measures (terracing, agroforestry and perennial herbaceous fodder). Assessment of the effectiveness of these local measures, combined with the RUSLE calculations, revealed that current average soil loss could be reduced through low-cost measures such as contouring (by 11%), fodder plants (by 16%), and drainage ditches (by 53%). More expensive measures such as terracing and agroforestry can reduce erosion by as much as 63% (for agroforestry) and 93% (for agroforestry combined with terracing). Indeed, scenario runs for different levels of tolerable erosion rates showed that more cost-intensive and technologically advanced measures would lead to greater reduction of soil loss. However, given economic conditions in Tajikistan, it seems advisable to support the spread of low-cost and labourextensive measures.