A complete and efficient CUDA-sharing solution for HPC clusters

In this paper we detail the key features, architectural design, and implementation of rCUDA, an advanced framework to enable remote and transparent GPGPU acceleration in HPC clusters. rCUDA allows decoupling GPUs from nodes, forming pools of shared accelerators, which brings enhanced flexibility to cluster configurations. This opens the door to configurations with fewer accelerators than nodes, as well as permits a single node to exploit the whole set of GPUs installed in the cluster. In our proposal, CUDA applications can seamlessly interact with any GPU in the cluster, independently of its physical location. Thus, GPUs can be either distributed among compute nodes or concentrated in dedicated GPGPU servers, depending on the cluster administrator’s policy. This proposal leads to savings not only in space but also in energy, acquisition, and maintenance costs. The performance evaluation in this paper with a series of benchmarks and a production application clearly demonstrates the viability of this proposal. Concretely, experiments with the matrix–matrix product reveal excellent performance compared with regular executions on the local GPU; on a much more complex application, the GPU-accelerated LAMMPS, we attain up to 11x speedup employing 8 remote accelerators from a single node with respect to a 12-core CPU-only execution. GPGPU service interaction in compute nodes, remote acceleration in dedicated GPGPU servers, and data transfer performance of similar GPU virtualization frameworks are also evaluated.     

Holistic schedulability analysis for multipacket messages in AFDX networks

The ARINC-664, Part 7 (AFDX) standard defines a communication network based on Ethernet and the UDP/IP protocols. Contrary to general-purpose Ethernet, the timing behavior in AFDX is deterministic due to the use of special network switches and end systems with static routing tables and traffic policing at the sending end through mechanisms called virtual links. Even though the latencies in this network are bounded, there are scheduling and contention effects that need to be analyzed. In this paper we develop a response-time analysis for multipacket messages transmitted through an AFDX network including the scheduling of the virtual links and sub-virtual links, and also the contention in the end systems and in the switches. This analysis allows us to obtain worst-case latencies and output jitter for the network messages with a precise modeling of the sending and receiving ends. These results can be integrated in a holistic approach with the response time analysis of the threads in the processing nodes to obtain end-to-end response times in heterogeneous distributed systems.     

Carbohydrate type and amount alter intravascular processing and catabolism of plasma lipoproteins in guinea pigs

To test the effects of exchanging dietary complex and simple carbohydrate for fat calories on lipoprotein metabolism, guinea pigs were fed two different fat/carbohydrate ratios: 2.5∶58% (w/w) or 25∶29% (w/w) with either sucrose or starch as the carbohydrate source. Animals fed high-fat had higher plasma low-density lipoprotein (LDL) and hepatic cholesterol concentrations than animals fed low-fat diets (P<0.01). The cholesteryl ester content per particle was higher, and the number of triacylglycerol (TAG) molecules was lower in very low density lipoprotein (VLDL) and LDL from animals fed high-fat diets. Intake of high-fat/sucrose resulted in higher plasma LDL concentrations than intake of high-fat/starch, and animals fed low-fat/starch had the highest plasma TAG concentrations associated with VLDL particles containing more TAG molecules, as well as a TAG-enriched LDL. The activity of plasma lecithin cholesteryl:acyl transferase (LCAT) was highest in animals fed high-fat/sucrose, and heart lipoprotein lipase (LPL) activity was higher in animals fed high-fat diets. Hepatic apoprotein B/E (apo B/E) receptor number (B_max) was increased 21% with low-fat diets (P<0.01). These results suggest that the hypercholesterolemia induced by high-fat and by sucrose intake are associated with a higher plasma LCAT activity which results in a cholesteryl ester-enriched VLDL which, by the action of LPL, might be more readily converted to LDL through the delipidation cascade leading to downregulation of hepatic apo B/E receptors. The hypertriglyceridemia associated with low-fat intake may result from increased production of VLDL TAG, which would explain the increased TAG content and the higher TAG/CE ratio of VLDL from animals fed the low-fat/starch diet.     

Photoluminiscent Manganese Nanoparticles from Solid State Polyphosphazenes Organometallic Derivatives

Pyrolysis in air at 800°C of [{NP(OC₁₂H₁₀)}_0.6{NP(OC₆H₄PPh₂·(π-CH₃C₅H₄)−Mn(CO)₂)₂}_0.4] _n (1) in the solid state affords product 2 containing nanoclusters of Mn₂P₂O₇ with sizes ranging from 50 to 90 nm and averaging about 74 nm. The egg-shape of the unpyrolyzed organometallic polymer is retained but with increased particle size after pyrolysis. The pyrolytic material shows near-infrared photoluminescence attributed to the emission of tetrahedral Mn²⁺ embedded in a matrix of Mn₂P₂O₇. The solid-state pyrolysis of organometallic derivatives of polyphosphazenes may be a useful and general route to nano-structured Mn₂P₂O₇. An erratum to this article can be found at     

I3WSN: Industrial Intelligent Wireless Sensor Networks for indoor environments

Sensor Web Enablement (SWE) technologies have been successfully applied to a great variety of outdoor scenarios but, in practical terms, little effort has been applied for indoor environments, and even less in the field of industrial applications. This article presents an intelligent SWE application for an indoor and industrial scenario, with the aim of improving and increasing the levels of human safety. The base low-level architecture is built on top of wireless sensor networks (WSN) connected to a Sensor Observation Service (SOS). Higher layers in the architecture include services that make real time decisions based on the collected data. Both simulation and experimental results are presented. The paper shows the viability of our approach in terms of performance, scalability, modularity and safety.     

Sustainability in Web server systems

The exponential growth of Internet during the last decade leads us to make more efforts in the researching and developing of sustainable Web servers in order to decrease the global energy demand. In this paper, we cover, as a first step, a proper review of the literature related to the energy efficiency research in Web server systems to depict the state of the art to plan further contributions as more research in sustainable Web systems. We also propose and implement an energy metric that permits to establish a relation between the Quality of Service (QoS) obtained by the system and the power it consumes.     

Study of the thermomechanical and electrical properties of conducting composites containing natural rubber and carbon black

This work describes the preparation and characterization of composite materials obtained by the combination of natural rubber (NR) and carbon black (CB) in different percentages, aiming to improve their mechanical properties, processability, and electrical conductivity, aiming future applications as transducer in pressure sensors. The composites NR/CB were characterized through optical microscopy (OM), DC conductivity, thermal analysis using differential scanning calorimetry (DSC), dynamic mechanical thermal analysis (DMA), thermogravimetry (TGA), and stress–strain test. The electrical conductivity varied between 10^?9 and 10 S m^?1, depending on the percentage of CB in the composite. Furthermore, a linear (and reversible) dependence of the conductivity on the applied pressure between 0 and 1.6 MPa was observed for the sample with containing 80 wt % of NR and 20% of CB. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci, 2007     

A web-based platform for biosignal visualization and annotation

With the advent of wearable sensing and mobile technologies, biosignals have seen an increasingly growing number of application areas, leading to the collection of large volumes of data. One of the difficulties in dealing with these data sets, and in the development of automated machine learning systems which use them as input, is the lack of reliable ground truth information. In this paper we present a new web-based platform for visualization, retrieval and annotation of biosignals by non-technical users, aimed at improving the process of ground truth collection for biomedical applications. Moreover, a novel extendable and scalable data representation model and persistency framework is presented. The results of the experimental evaluation with possible users has further confirmed the potential of the presented framework.     

Beast methodology: An agile testing methodology for multi-agent systems based on behaviour driven development

This paper presents a testing methodology to apply Behaviour Driven Development (BDD) techniques while developing Multi-Agent Systems (MASs), termed BEhavioural Agent Simple Testing (BEAST) Methodology. This methodology is supported by the open source framework (BEAST Tool) which automatically generates test cases skeletons from BDD scenarios specifications. The developed framework allows the testing of MASs based on JADE or JADEX platforms. In addition, this framework offers a set of configurable Mock Agents with the aim of being able to execute tests while the MAS is under development. The BEAST Methodology presents transparent traceability from user requirements to test cases. Thus, the stakeholders can be aware of the project status. The methodology and the associated tool have been validated in the development of a MAS for fault diagnosis in FTTH (Fiber To The Home) networks. The results have been measured in quantifiable way obtaining a reduction of the tests implementation time.     

A genetic algorithm for job-shop scheduling with operators enhanced by weak Lamarckian evolution and search space narrowing

The job-shop scheduling problem with operators is a very interesting problem that generalizes the classic job-shop problem in such a way that an operation must be algorithm to solve this problem considering makespan minimization. The genetic algorithm uses permutations with repetition to encode chromosomes and a schedule generation scheme, termed OG&T, as decoding algorithm. This combination guaranties that at least one of the chromosomes represents and optimal schedule and, at the samhat machines and operators are idle while an operation is available to be processed. To improve the quality of the schedules for large instances, we use Lamarckian evolution and modify the OG&T algorithm to further reduce the idle time of the machines and operators, in this case at the risk of leaving all optimal schedules out of the search space. We conducted a large experimental study showing that these improvements allow the genetic algorithm to reach high quality solutions in very short time, and so it is quite competitive with the current state-of-the-art methods.