Insights into the Fallback Path of Best-Effort Hardware Transactional Memory Systems

Current industry proposals for Hardware Transactional Memory (HTM) focus on best-effort solutions (BE-HTM) where hardware limits are imposed on transactions. These designs may show a significant performance degradation due to high contention scenarios and different hardware and operating system limitations that abort transactions, e.g. cache overflows, hardware and software exceptions, etc. To deal with these events and to ensure forward progress, BE-HTM systems usually provide a software fallback path to execute a lock-based version of the code. In this paper, we propose a hardware implementation of an irrevocability mechanism as an alternative to the software fallback path to gain insight into the hardware improvements that could enhance the execution of such a fallback. Our mechanism anticipates the abort that causes the transaction serialization, and stalls other transactions in the system so that transactional work loss is mini- mized. In addition, we evaluate the main software fallback path approaches and propose the use of ticket locks that hold precise information of the number of transactions waiting to enter the fallback. Thus, the separation of transactional and fallback execution can be achieved in a precise manner. The evaluation is carried out using the Simics/GEMS simulator and the complete range of STAMP transactional suite benchmarks. We obtain significant performance benefits of around twice the speedup and an abort reduction of 50% over the software fallback path for a number of benchmarks.

Furie, curiosité, solidarité : mise en scène de la tension dans le cinéma des frères Dardenne

Ce mémoire de recherche-création est accompagné du court métrage «Tala». Pour visionner en ligne : vimeo.com/ondemand/talapierphilippechevigny

Furie, curiosité, solidarité : mise en scène de la tension dans le cinéma des frères Dardenne

Ce mémoire de recherche-création est accompagné du court métrage «Tala». Pour visionner en ligne : vimeo.com/ondemand/talapierphilippechevigny

Proposal to minimize the difficulty of recording abortion data in sheep

Background: The most frequent viral diseases which can cause abortion in sheep are Blue tongue, Border disease virus, Cache Valley fever and Schmallenberg virus. The diagnosis of abortion, namely virus-induced represents a challenge to field clinicians, since clinical signs presented by the dam are discrete, non-specific and variable (Agerhom et al., 2015). On the other hand, while some foetuses reveal characteristic and visible malformations, others do not reveal any lesions. In face of it, definitive diagnosis requires an appropriate history collection, as well as sending fresh samples, namely abortion material, foetus, placenta and umbilical cord, to a specialty laboratory, to obtain a precise diagnosis. Objectives: The authors suggest a registration method of all mandatory data, in order to further assist the diagnosis of viral diseases at the laboratories, including the most frequent congenital malformations reported in sheep abortions. Methods: Abortion samples of suspected viral origin were collected and all data were registered, in worktables optimized for this purpose. Results: The authors document, using macroscopic figures lesions of malformations in abortions, emphasizing the frequency and the importance of documenting each case, proposing practical and effective worktables to assist the fieldwork. Conclusions: Field clinician’s awareness of the importance of early detection of viral diseases causing abortion outbreaks stimulates a proper data collection for each case of abortion, in order to contribute to a precise diagnosis and posterior consistent epidemiological studies, which may allow diminishing of economic losses.

Die Rolle der Internet Service Provider im deutschen und US-amerikanischen Recht - Verantwortlichkeit und Privilegierung für Urheberrechtsverletzungen im Internet

Die Dissertation befasst sich mit der Verantwortlichkeit und der Haftungsprivilegierung von Internet Service Providern (ISP) für Urheberrechtsverletzungen Dritter im deutschen und US-amerikanischen Recht. Internet Service Provider (das sind Host-Provider, Cache-Provider, Access-Provider sowie Suchmaschinen-Anbieter bzw. sonstige Linksetzende) nehmen eine wichtige Rolle im Internet ein. In Deutschland sehen sich die ISP jedoch trotz gesetzlicher Haftungsprivilegien erheblichen Rechtsunsicherheiten ausgesetzt. Der Dissertation liegt die Hypothese zugrunde, dass durch die Ausweitung der Verantwortlichkeit der ISP durch die Rechtsprechung die gesetzlich stipulierte Haftungsprivilegierung faktisch entwertet wird. Insbesondere stehen die Nichtanwendung der Privilegien auf Unterlassungsansprüche sowie die im Rahmen der Störerhaftung begründeten Prüfpflichten der Intention des Gesetzgebers (auch auf europäischer Ebene) entgegen. Eine gerechte Balance der Interessen der Akteure durch die Rechtsprechung wurde nicht erreicht. Die gesetzliche Ausgestaltung der US-amerikanischen Privilegierung verspricht hingegen eine den Interessen der ISP gerechtere Lösung. Auch die Rechtsanwendung der US-amerikanischen Gerichte liegt hiermit auf einer Linie.

Análisis económico de los bienes baldíos en Colombia

En este estudio se pretende exponer la situación actual de los derechos de propiedad de los bienes inmuebles rurales, principalmente bienes baldíos. Se parte de la base de los regímenes que han existido, sus principales características, su estado actual, la productividad económica de los mismos en determinadas regiones de Colombia, su seguridad jurídica y se planteará una propuesta de mejoramiento que podría obtener mejores resultados, tanto sociales como económicos, soportado en la experiencia del autor y en los diversos estudios que se han efectuado sobre el tema a nivel mundial.

Actitudes hacia la Enfermedad del Alzheimer según el género y la edad de adultos colombianos

La enfermedad de Alzheimer (EA) es la demencia más frecuente y su prevalencia continúa en aumento tanto en Colombia como en el mundo. Esta investigación tuvo como objetivo explorar si las actitudes hacia la EA varían según la edad y género de 450 personas adultas colombianas. Se realizó un estudio exploratorio de corte transversal en el que se aplicó un cuestionario autodiligenciado. Se encontró que efectivamente hay algunas diferencias según la edad y el género en el componente cognoscitivo (creencias y conocimiento) y conductual (intención conductual y conducta) de las actitudes; y diferencias según el género en el componente afectivo. Se concluye que los conocimientos sobre la EA son escasos, que la tristeza es la emoción predominante hacia la EA y que es un tema de interés en el que predomina la idea de que afecta especialmente la memoria. Se discutieron los resultados reconociendo que esta es una aproximación inicial a las actitudes hacia la EA.

Calibração radiométrica de instrumentação óptica espectrofotométrica

Os espectrómetros instalados a bordo de satélites e à superfície da Terra têm desempenhado um papel fundamental na compreensão da química e dinâmica da atmosfera e na monitorização da poluição ambiental. O SPATRAM (SPectrometer for Atmosferic TRAcers Measurements) é um espectrómetro ultravioleta – Visível que compreende a região espectral entre 250-950nm e se encontra instalado no Instituto de Ciências da Terra (ICT) desde Abril de 2004. Enquanto isso, em 2012, um novo modelo do instrumento SPATRAM foi desenvolvido no ICT e foi chamado SPATRAM2. O objectivo do trabalho proposto é a calibração radiométrica do espectrómetro SPATRAM2, utilizando uma lâmpada de halogéneo e uma esfera de integração. A calibração radiométrica do sistema SPATRAM2 permitirá obter a radiação solar directa, com alta resolução espectral, o que actualmente não se encontra disponível. Este tipo de medição poderá ter um papel importante na investigação e desenvolvimento na área da energia solar e aplicações; Radiometric Calibration of Spectrophotometric Optical Instrumentation Abstract: Spectrometers installed aboard satellites and located on Earth’s surface have played a fundamental role to understand atmosphere’s chemistry and dynamic and to monitor environmental pollution. The SPATRAM (SPectrometer for Atmosferic TRAcers Measurements) instrument is an ultraviolet spectrometer – visible that covers spectral region between 250-950nm and it is installed in Instituto de Ciências da Terra (Institute of Earth Sciencies), ICT, since April 2004. Meanwhile, in 2012, a new model of SPATRAM instrument was developed in ICT and was called SPATRAM2. The goal of this project is the radiometric calibration of the SPATRAM2 spectrometer using a halogen lamp and an integrating sphere. Radiometric calibration of SPATRAM2 system will provide direct solar radiation, with high spectral resolution, that is not available nowadays. This type of measurement may play an important role in solar energy’s progress and investigation.

A first survey on the status of the Posidonia oceanica meadow in Calpe Bay (Alicante, Spain).

Posidonia oceanica, endemic seagrass of the Mediterranean Sea, forms extensive meadows. It is included among the Mediterranean protected habitats by the Habitat Directive (92/43/EEC). P. oceanica meadows are exposed to anthropogenic impacts that are more evident in areas close to cities, ports or areas with a large coastal tourism development. Mean exponential decline rate of 5 % yr-1 is estimated for the Spanish meadows. If this trend is maintained, most of the meadows are predicted to halve in shoot density over the next 20 years. The meadows regression can give way to a new regime, which supposes the loss of the multiple services that the meadows provided. It is necessary to recognize situations of stress in time, before irreversible damages and changes towards alternative regimes are evident. This study has been carried out in Calpe Bay, Alicante (Spain), during May and June 2017, with the aim of assessing, for the first time, the status of the P. oceanica meadows providing a baseline data for the future monitoring scheme. The features and status of the seagrass beds have been assessed by physical, physiographical, structural and functional descriptors. The results showed that the health status classification of P. oceanica meadows in Calpe Bay vary between “equilibrium” and “disturbed”. The “disturbed” conditions were observed in a shaded area where it is probably due to the low solar radiance. In a lower limit in a shallow meadow, where it could be due to the combined effect of substrate structure and hydrodynamic regime. Finally in a touristic area where patchy impacts could be attributed to direct human disturbance (e.g. anchoring). Overall the status of P. oceanica meadows in Calpe bay is not worrying. However, it is important to develop monitoring plans to assess the dynamics of the seagrass detecting any early decline symptom in order to act, as soon as possible because, when a regression of a meadow is produced, it could not be recovered at human scales.

Deploying and processing neural representations of signals

Neural representations (NR) have emerged in the last few years as a powerful tool to represent signals from several domains, such as images, 3D shapes, or audio. Indeed, deep neural networks have been shown capable of approximating continuous functions that describe a given signal with theoretical infinite resolution. This finding allows obtaining representations whose memory footprint is fixed and decoupled from the resolution at which the underlying signal can be sampled, something that is not possible with traditional discrete representations, e.g., grids of pixels for images or voxels for 3D shapes. During the last two years, many techniques have been proposed to improve the capability of NR to approximate high-frequency details and to make the optimization procedures required to obtain NR less demanding both in terms of time and data requirements, motivating many researchers to deploy NR as the main form of data representation for complex pipelines. Following this line of research, we first show that NR can approximate precisely Unsigned Distance Functions, providing an effective way to represent garments that feature open 3D surfaces and unknown topology. Then, we present a pipeline to obtain in a few minutes a compact Neural Twin® for a given object, by exploiting the recent advances in modeling neural radiance fields. Furthermore, we move a step in the direction of adopting NR as a standalone representation, by considering the possibility of performing downstream tasks by processing directly the NR weights. We first show that deep neural networks can be compressed into compact latent codes. Then, we show how this technique can be exploited to perform deep learning on implicit neural representations (INR) of 3D shapes, by only looking at the weights of the networks.

Smart sampling approaches for Decision-Focused Learning

Many real-word decision- making problems are defined based on forecast parameters: for example, one may plan an urban route by relying on traffic predictions. In these cases, the conventional approach consists in training a predictor and then solving an optimization problem. This may be problematic since mistakes made by the predictor may trick the optimizer into taking dramatically wrong decisions. Recently, the field of Decision-Focused Learning overcomes this limitation by merging the two stages at training time, so that predictions are rewarded and penalized based on their outcome in the optimization problem. There are however still significant challenges toward a widespread adoption of the method, mostly related to the limitation in terms of generality and scalability. One possible solution for dealing with the second problem is introducing a caching-based approach, to speed up the training process. This project aims to investigate these techniques, in order to reduce even more, the solver calls. For each considered method, we designed a particular smart sampling approach, based on their characteristics. In the case of the SPO method, we ended up discovering that it is only necessary to initialize the cache with only several solutions; those needed to filter the elements that we still need to properly learn. For the Blackbox method, we designed a smart sampling approach, based on inferred solutions.