Parallel implementation of a robust optical flow technique

[EN] The accuracy and performance of current variational optical ow methods have considerably increased during the last years. The complexity of these techniques is high and enough care has to be taken for the implementation. The aim of this work is to present a comprehensible implementation of recent variational optical flow methods. We start with an energy model that relies on brightness and gradient constancy terms and a ow-based smoothness term. We minimize this energy model and derive an e cient implicit numerical scheme. In the experimental results, we evaluate the accuracy and performance of this implementation with the Middlebury benchmark database. We show that it is a competitive solution with respect to current methods in the literature. In order to increase the performance, we use a simple strategy to parallelize the execution on multi-core processors.

A parallel future for Ocean research?

[EN] Today, science is difficult to pursue because funding is so tenuous. In such a financial climate, researchers need to consider parallel alternatives to ensure that scientific research can continue. Based on this thinking, we created BIOCEANSolutions, a company born of a research group. A great variety of environmental regulations and standards have emerged over recent years with the purpose of protecting natural ecosystems. These have enabled us to link our research to the market of environmental management. Marine activities can alter environmental conditions, resulting in changes in physiological states, species diversity, abundance, and biomass in the local biological communities. In this way, we can apply our knowledge, to plankton ecophysiology and biochemical oceanography. We measure enzyme activities as bio-indicators of energy metabolism and other physiological rates and biologic-oceanographic processes in marine organisms. This information provides insight into the health of marine communities, the stress levels of individual organisms, and potential anomalies that may be affecting them. In the process of verifying standards and complying with regulations, we can apply our analytic capability and knowledge. The main analyses that we offer are: (1) the activity of the electron transport system (ETS) or potential respiration (Φ), (2) the physiological measurement of respiration (oxygen consumption), (3) the activity of Isocitrate dehydrogenase (IDH), (4) the respiratory CO2 production, and (5) the activity of Glutamate dehydrogenase (GDH) and (6) the physiological measurement of ammonium excretion. In addition, our experience in a productive research group allows us to pursue and develop technical-experimental activities such as marine and freshwater aquaculture, oceanographic field sampling, as well as providing guidance, counseling, and academic services. In summary, this new company will permit us to create a symbiosis between public and private sectors that serve clients and will allow us to grow and expand as a research team.

Estudio de utilización efectiva de procesadores vectoriales

[ES] La vectorización es un proceso de explotación de paralelismo de datos muy potente que, bien usado permite obtener un mejor rendimiento de la ejecución de las aplicaciones. Debido a ello, hoy en día muchos procesadores incluyen extensiones vectoriales en su repositorio de instrucciones. Para las máquinas basadas en estos procesadores, existen multitud de compiladores que permiten explotar la vectorización. Sin embargo, no todas las aplicaciones experimentan una mejora en el rendimiento cuando son vectorizadas, y no todos los compiladores son capaces de extraer el mismo rendimiento vectorial de las aplicaciones. Este trabajo presenta un estudio exhaustivo del rendimiento de diversas aplicaciones numéricas, con el objetivo de determinar el grado de utilización efectiva de la unidad vectorial. Tras seleccionar los benchmarks Polyhedron, Mantevo, Sequoia, SPECfp y NPB, se compilaron activando la vectorización y se simularon en una versión modificada del simulador de cache CMPSim, enriquecida con un núcleo basado en el coprocesador Intel Xeon Phitm. En aquellos casos en que la utilización era baja, se realizó un diagnóstico a nivel de software de la fuente del problema y se propusieron mejoras que podrían aumentar el uso efectivo de la unidad vectorial. Para aquellas aplicaciones limitadas por memoria, se realizó un diagnóstico a nivel de hardware con el fin de determinar hasta que punto el diseño de la máquina repercute en el rendimiento de la aplicación en casos de buen uso de la unidad vectorial.

Performance evaluation of a parallel algorithm for simultaneous untangling and smoothing of tetrahedral meshes

[EN]A new parallel algorithm for simultaneous untangling and smoothing of tetrahedral meshes is proposed in this paper. We provide a detailed analysis of its performance on shared-memory many-core computer architectures. This performance analysis includes the evaluation of execution time, parallel scalability, load balancing, and parallelism bottlenecks. Additionally, we compare the impact of three previously published graph coloring procedures on the performance of our parallel algorithm. We use six benchmark meshes with a wide range of sizes. Using these experimental data sets, we describe the behavior of the parallel algorithm for different data sizes. We demonstrate that this algorithm is highly scalable when it runs on two different high-performance many-core computers with up to 128 processors...