Parallel Integration Research Articles

Heuristic parallel selective ensemble algorithm based on clustering and improved simulated annealing

Ensemble learning can achieve stronger generalization ability by training multiple weak classifier models and fusing the prediction results of these weak classifiers. However, with the increase in the number of weak classifier models, the complexity and training time of the model are also increasing. This paper introduces a heuristic parallel selective ensemble algorithm based on clustering and improved simulated annealing. First, the method uses selective clustering to remove similar weak classifiers, so as to reduce the number of weak models, and select the weak classifiers with large differences as candidate sets. Then, the optimal classifier sequence set is selected based on the improved simulated annealing heuristic selective ensemble algorithm, which improves the classification performance of the model. Moreover, the parallel integration strategy is adopted to improve the efficiency of selecting optimal subset of classifiers, thus effectively reducing the training time of the model. Experimental results show that the proposed algorithm exhibits better performance compared with traditional methods.

GinSODA: massive parallel integration of stiff ODE systems on GPUs

Ordinary differential equations (ODEs) are a widespread formalism for the mathematical modeling of natural and engineering systems, whose analysis is generally performed by means of numerical integration methods. However, real-world models are often characterized by stiffness, a circumstance that can lead to prohibitive execution times. In such cases, the practical viability of many computational tools—e.g., sensitivity analysis—is hampered by the necessity to carry out a large number of simulations. In this work, we present ginSODA, a general-purpose black-box numerical integrator that distributes the calculations on graphics processing units, and allows to run massive numbers of numerical integrations of ODE systems characterized by stiffness. By leveraging symbolic differentiation, meta-programming techniques, and source code hashing, ginSODA automatically builds highly optimized binaries for the CUDA architecture, preventing code re-compilation and allowing to speed up the computation with respect to the sequential execution. ginSODA also provides a simplified Python interface, which allows to define a system of ODEs and the test to be performed in a few lines of code. According to our results, ginSODA provides up to a $$25 \times $$ speedup with respect to the sequential execution.

A Test Platform for the Noise Characterization of SiGe Microbolometer ROICs

This paper introduces an in-circuit performance evaluation system for SiGe microbolometer readout integrated circuits (ROICs) that can characterize the overall system noise performance by emulating microbolometers with MOSFETs biased in the triode region. Specifically, the proposed test platform is designed for the testing of imagers with high resistance SiGe microbolometers. The architecture of the ROIC is based on a bridge with active and reference bolometer pixels with a capacitive transimpedance amplifier input stage and column parallel integration with serial readout. Noise measurements along with simulated resistance curves of the dummy detectors are reported. The prototype with 17-μm pixel pitch has been designed and fabricated in a 0.25-μm SiGe BiCMOS process.

Nanoscale investigation of the silicon carbide double-diffused MOSFET with scanning capacitance force microscopy

Power semiconductor devices are progressing towards high breakdown voltages and are fabricated using wide-band-gap semiconductor materials and parallel integration facilitated by microfabrication techniques. We achieved nanoscale observation of power semiconductor devices by scanning probe microscopy (SPM) based on the combination of atomic force microscopy (AFM), Kelvin probe force microscopy (KFM), and scanning capacitance force microscopy (SCFM), which enabled direct observation of internal structures and channel formation in a silicon carbide double-diffused MOSFET at nanoscale resolution. Nanoscale observations were performed with stability control using the frequency-modulation (FM) detection method in a vacuum-pressure environment, with and without bias voltage applied to the MOSFET.

Electrochemically Scalable Production of Fluorine-Modified Graphene for Flexible and High-Energy Ionogel-Based Microsupercapacitors.

Scalable production of high-quality heteroatom-modified graphene is critical for microscale supercapacitors but remains a great challenge. Herein, we demonstrate a scalable, single-step electrochemical exfoliation of graphite into highly solution-processable fluorine-modified graphene (FG), achieved in an aqueous fluorine-containing neutral electrolyte, for flexible and high-energy-density ionogel-based microsupercapacitors (FG-MSCs). The electrochemically exfoliated FG nanosheets are characterized by atomic thinness, large lateral size (up to 12 μm), a high yield of >70% with ≤3 layers, and a fluorine doping of 3 at%, allowing for large-scale production of FG-MSCs. Our ionogel-based FG-MSCs deliver high energy density of 56 mWh cm-3, by far outperforming the most reported MSCs. Furthermore, the all-solid-state microdevices offer exceptional cyclability with ∼93% after 5000 cycles, robust mechanical flexibility with 100% of capacitance retention bended at 180°, and outstanding serial and parallel integration without the requirement of metal-based interconnects for high-voltage and high-capacitance output. Therefore, these FG-MSCs represent remarkable potential for electronics.

Conceptual design and preliminary performance analysis of a hybrid nuclear-solar power system with molten-salt packed-bed thermal energy storage for on-demand power supply

In support of more efficient utilization of solar and nuclear energy in power generation, the present work proposes a conceptual design of a hybrid nuclear-solar power system (HNSPS) for on-demand power supply, based on a parallel thermal integration of small modular reactors with commercialized molten-salt concentrating solar power tower plants. A cost-effective thermal energy storage system is employed as a thermal coupling and a heat dispatcher to coordinate both nuclear heat and solar heat with power demands. The overall configuration and operation strategy of the hybrid system are elaborated. A numerical model of the hybrid system is developed based on the circulation of molten-salt. Using the model, a 7-day performance analysis of the hybrid system with different considerations are conducted. Effects of key design parameters on system performances are investigated by a parametric study. A customized evaluation parameter, named as “design score”, is defined to assess the suitability of the designed configurations. The performance analysis shows that the integration of nuclear power with concentrating solar power allows the system to be less dependent on heat dispatch to achieve required power regulations, comparing to a standalone CSP plant with an equivalent power capacity. The parametric study indicates that a higher heat generation utilization of the hybrid system requires a larger TES and a smaller heliostats field, while a higher power supply efficiency of the hybrid system is accompanied with a larger TES, a larger heliostats field, and a higher nuclear power capacity proportion. The optimum design configuration of a 200 MWe HNSPS is estimated to have a nuclear power capacity proportion of 50%, a theoretical thermal energy storage duration of 14.8 h, and a heliostats field with a solar multiple of 1.27. This configuration can completely satisfy the power demand without any heat generation curtailment during the 7-day operation. The obtained results allow the HNSPS to be considered as a promising non-carbon-emitting power generation system for on-demand power supply.

Transfer-printing-based integration of a III-V-on-silicon distributed feedback laser.

An electrically pumped DFB laser integrated on and coupled to a silicon waveguide circuit is demonstrated by transfer printing a 40 × 970 μm2 III-V coupon, defined on a III-V epitaxial wafer. A second-order grating defined in the silicon device layer with a period of 477 nm and a duty cycle of 75% was used for realizing single mode emission, while an adiabatic taper structure is used for coupling to the silicon waveguide layer. 18 mA threshold current and a maximum single-sided waveguide-coupled output power above 2 mW is obtained at 20°C. Single mode operation around 1550 nm with > 40 dB side mode suppression ratio (SMSR) is realized. This new integration approach allows for the very efficient use of the III-V material and the massively parallel integration of these coupons on a silicon photonic integrated circuit wafer. It also allows for the intimate integration of III-V opto-electronic components based on different epitaxial layer structures.

Factoring Integers With a Brain-Inspired Computer

The bound to factor large integers is dominated by the computational effort to discover numbers that are B-smooth, i.e., integers whose largest prime factor does not exceed B. Smooth numbers are traditionally discovered by sieving a polynomial sequence, whereby the logarithmic sum of prime factors of each polynomial value is compared to a threshold. On a von Neumann architecture, this requires a large block of memory for the sieving interval and frequent memory updates, resulting in O(ln ln B) amortized time complexity to check each value for smoothness. This paper presents a neuromorphic sieve that achieves a constant-time check for smoothness by reversing the roles of space and time from the von Neumann architecture and exploiting two characteristic properties of brain-inspired computation: massive parallelism and constant time synaptic integration. The effects on sieving performance of two common neuromorphic architectural constraints are examined: limited synaptic weight resolution, which forces the factor base to be quantized, and maximum indegree, which forces the factor base to be partitioned. Strategies for dealing with both constraints are introduced and the approach is validated by modifying msieve , which implements the multiple polynomial quadratic sieve, to use the IBM Neurosynaptic System as a coprocessor for integer factorization.

A parallel time integrator for noisy nonlinear oscillatory systems

In this paper, we adapt a parallel time integration scheme to track the trajectories of noisy non-linear dynamical systems. Specifically, we formulate a parallel algorithm to generate the sample path of nonlinear oscillator defined by stochastic differential equations (SDEs) using the so-called parareal method for ordinary differential equations (ODEs). The presence of Wiener process in SDEs causes difficulties in the direct application of any numerical integration techniques of ODEs including the parareal algorithm. The parallel implementation of the algorithm involves two SDEs solvers, namely a fine-level scheme to integrate the system in parallel and a coarse-level scheme to generate and correct the required initial conditions to start the fine-level integrators. For the numerical illustration, a randomly excited Duffing oscillator is investigated in order to study the performance of the stochastic parallel algorithm with respect to a range of system parameters. The distributed implementation of the algorithm exploits Massage Passing Interface (MPI).

Photo-Induced Room-Temperature Gas Sensing with a-IGZO Based Thin-Film Transistors Fabricated on Flexible Plastic Foil.

We present a gas sensitive thin-film transistor (TFT) based on an amorphous Indium–Gallium–Zinc–Oxide (a-IGZO) semiconductor as the sensing layer, which is fabricated on a free-standing flexible polyimide foil. The photo-induced sensor response to NO2 gas at room temperature and the cross-sensitivity to humidity are investigated. We combine the advantages of a transistor based sensor with flexible electronics technology to demonstrate the first flexible a-IGZO based gas sensitive TFT. Since flexible plastic substrates prohibit the use of high operating temperatures, the charge generation is promoted with the help of UV-light absorption, which ultimately triggers the reversible chemical reaction with the trace gas. Furthermore, the device fabrication process flow can be directly implemented in standard TFT technology, allowing for the parallel integration of the sensor and analog or logical circuits.

Диаспоростроительство на Урале: от «переселенческих обществ» до трансграничных мигрантов

In the formation of the ethno-cultural landscape of the Urals, the processes of integration of representatives of migrant, religious and ethnocultural minorities played an important role. In parallel, both individual and group integration mechanisms were applied. It is advisable to consider group adaptation strategies as a process of diaspora construction. The experience of the formation and development of diaspora groups clearly demonstrates the possibility of various cultural grounds for this process, both socially-confessional and actually ethnic. The presented article summarizes the dynamics of diaspora-building in the Urals, formulated the main problem fields for further research. Collective strategies for the integration of minorities have demonstrated that the diaspora is not a given. First and foremost, this is a specific way of life, the organization of human relations and ties. Diasporality demonstrates not only a huge variety of identification and institutional options, but also an amazing dynamics of internal transformation. In the middle of the XIX — beginning of the XXI centuries. migrant have become an integral part of the ethno-cultural mosaic of the Urals. The diaspora strategy and adaptation practice was based on social institutions and life practices. They were based on the idea of the unity of fate and the place of exodus, the value of memory about the «country of exodus' or historical homeland'.

Parallel Integral Equation Based Nonoverlapping DDM for Solving Electrically-Large Scattering Problems in Half Space

In this paper, a parallel nonoverlapping non-conformal domain decomposition method using combined-field integral equation is proposed for fast and accurate analysis of electrically large objects situated in half space. This method decomposes the whole model into several easily solvable closed subdomains in a flexible way, and an explicit boundary condition is applied to ensure the continuity of electric currents across the boundary. The adaptive direction partitioning parallelization scheme for the oct-tree of a multilevel fast multipole algorithm is adopted to accelerate matrix-vector multiplications of subdomains as well as the coupling between them. The introducing of real image source to account for far interaction matrix greatly accelerates the process and reduces memory requirement with precision guaranteed. Numerical examples demonstrate that the proposed method is able to simulate realistic problems with a maximum dimension greater than 1000 wavelengths.

GraviDy, a GPU modular, parallel direct-summation N-body integrator: dynamics with softening

A wide variety of outstanding problems in astrophysics involve the motion of a large number of particles ($N\gtrsim 10^{6}$) under the force of gravity. These include the global evolution of globular clusters, tidal disruptions of stars by a massive black hole, the formation of protoplanets and the detection of sources of gravitational radiation. The direct-summation of $N$ gravitational forces is a complex problem with no analytical solution and can only be tackled with approximations and numerical methods. To this end, the Hermite scheme is a widely used integration method. With different numerical techniques and special-purpose hardware, it can be used to speed up the calculations. But these methods tend to be computationally slow and cumbersome to work with. Here we present a new GPU, direct-summation $N-$body integrator written from scratch and based on this scheme. This code has high modularity, allowing users to readily introduce new physics, it exploits available high-performance computing resources and will be maintained by public, regular updates. The code can be used in parallel on multiple CPUs and GPUs, with a considerable speed-up benefit. The single GPU version runs about 200 times faster compared to the single CPU version. A test run using 4 GPUs in parallel shows a speed up factor of about 3 as compared to the single GPU version. The conception and design of this first release is aimed at users with access to traditional parallel CPU clusters or computational nodes with one or a few GPU cards.

Competence Creep Revisited

AbstractHow is it that regardless of the reforms introduced by the Lisbon Treaty to better contain European integration in areas of core state powers, ‘competence creep’ can continue? What is the underlying cause? And why is it problematic? This article proposes answers to these questions through a systematic (re‐)conceptualization of the problem of ‘competence creep’, arguing that it results from the cross‐cutting governance that is the legal Leitmotif of European integration as well as from ‘two‐level games’ of national governments, and that it is problematic from the viewpoint of democratic legitimacy. However, it argues that the one form of competence creep that is most commonly understood as the core problem, and on which most reforms have focused, namely indirect legislation in areas of Member State competence, is actually the least worrying type of covert integration; negative and parallel integration, soft law and co‐ordination are all far more problematic.

Simplified fabrication of high areal capacitance all-solid-state micro-supercapacitors based on graphene and MnO2 nanosheets

All-solid-state micro-supercapacitors are acknowledged as a very promising class of microscale energy storage devices for directly integrating portable and wearable electronics. However, the improvement of electrochemical performance from materials to devices still remains tremendous challenges. Here, we demonstrate a novel and universal mask-assisted filtration technology for the simplified fabrication of all-solid-state planar micro-supercapacitors (MSCs) based on interdigital patterns of 2D pseudocapacitive MnO2 nanosheets and electrochemically exfoliated graphene film as both electrode and current collector, and polyvinyl alcohol/LiCl gel as electrolyte. Remarkably, the resulting MSCs exhibit outstanding areal capacitance of ∼355mF/cm2, which is among the highest values reported in the state-of-the-art MSCs. Meanwhile, MSCs possess exceptionally mechanical flexibility as high as ∼92% of initial capacitance even at a highly bending angle of 180°, excellent cyclability with a capacitance retention of 95% after 3000 cycles, and impressive serial or parallel integration for modulating the voltage or capacitance. Therefore, our proposed strategy of simplified construction of MSCs will pave the ways for utilizing graphene and analogous pseudocapactive nanosheets in high-performance MSCs.

Digoxin reveals a functional connection between HIV-1 integration preference and T-cell activation.

HIV-1 integrates more frequently into transcribed genes, however the biological significance of HIV-1 integration targeting has remained elusive. Using a selective high-throughput chemical screen, we discovered that the cardiac glycoside digoxin inhibits wild-type HIV-1 infection more potently than HIV-1 bearing a single point mutation (N74D) in the capsid protein. We confirmed that digoxin repressed viral gene expression by targeting the cellular Na+/K+ ATPase, but this did not explain its selectivity. Parallel RNAseq and integration mapping in infected cells demonstrated that digoxin inhibited expression of genes involved in T-cell activation and cell metabolism. Analysis of >400,000 unique integration sites showed that WT virus integrated more frequently than N74D mutant within or near genes susceptible to repression by digoxin and involved in T-cell activation and cell metabolism. Two main gene networks down-regulated by the drug were CD40L and CD38. Blocking CD40L by neutralizing antibodies selectively inhibited WT virus infection, phenocopying digoxin. Thus the selectivity of digoxin depends on a combination of integration targeting and repression of specific gene networks. The drug unmasked a functional connection between HIV-1 integration and T-cell activation. Our results suggest that HIV-1 evolved integration site selection to couple its early gene expression with the status of target CD4+ T-cells, which may affect latency and viral reactivation.

Skybridge-3D-CMOS: A Fine-Grained 3D CMOS Integrated Circuit Technology

Parallel and monolithic three-dimensional (3-D) integration directions realize 3-D integrated circuits (ICs) by utilizing layer-by-layer implementations, with each functional layer being composed in 2-D. In contrast, vertically composed 3-D CMOS has eluded us likely due to the seemingly insurmountable requirement of highly customized complex routing and regional 3-D doping to form and connect CMOS pull-up and pull-down networks in 3-D. In the current layer-by-layer directions, routing can be worse than 2D CMOS because of the limited pin access. In this paper, we propose Skybridge-3D-CMOS (S3DC), an IC fabric that shows for the first time a pathway to achieve fine-grained static CMOS circuit implementations using the vertical direction while also solving 3-D routability. It employs a new fabric assembly scheme based on predoped vertical nanowire bundles. It implements circuits in and across nanowires. It utilizes unique connectivity features to achieve CMOS connectivity in 3-D with excellent routability. As compared to the usually severely congested monolithic 3-D implementations, S3DC eliminates the routing congestions in all benchmarks studied. Further results, for the implemented benchmarks, show 56–77% reductions in power consumption, 4X–90X increases in density, and 20% loss to 9% benefit in best operating frequencies compared with the transistor-level monolithic 3-D technology.

Serial, parallel and hierarchical decision making in primates.

The study of decision-making has mainly focused on isolated decisions where choices are associated with motor actions. However, problem-solving often involves considering a hierarchy of sub-decisions. In a recent study (Lorteije et al. 2015), we reported behavioral and neuronal evidence for hierarchical decision making in a task with a small decision tree. We observed a first phase of parallel evidence integration for multiple sub-decisions, followed by a phase in which the overall strategy formed. It has been suggested that a 'flat' competition between the ultimate motor actions might also explain these results. A reanalysis of the data does not support the critical predictions of flat models. We also examined the time-course of decision making in other, related tasks and report conditions where evidence integration for successive decisions is decoupled, which excludes flat models. We conclude that the flexibility of decision-making implies that the strategies are genuinely hierarchical.

Dermoscopic Image Segmentation via Multistage Fully Convolutional Networks.

Segmentation of skin lesions is an important step in the automated computer aided diagnosis of melanoma. However, existing segmentation methods have a tendency to over- or under-segment the lesions and perform poorly when the lesions have fuzzy boundaries, low contrast with the background, inhomogeneous textures, or contain artifacts. Furthermore, the performance of these methods are heavily reliant on the appropriate tuning of a large number of parameters as well as the use of effective preprocessing techniques, such as illumination correction and hair removal. We propose to leverage fully convolutional networks (FCNs) to automatically segment the skin lesions. FCNs are a neural network architecture that achieves object detection by hierarchically combining low-level appearance information with high-level semantic information. We address the issue of FCN producing coarse segmentation boundaries for challenging skin lesions (e.g., those with fuzzy boundaries and/or low difference in the textures between the foreground and the background) through a multistage segmentation approach in which multiple FCNs learn complementary visual characteristics of different skin lesions; early stage FCNs learn coarse appearance and localization information while late-stage FCNs learn the subtle characteristics of the lesion boundaries. We also introduce a new parallel integration method to combine the complementary information derived from individual segmentation stages to achieve a final segmentation result that has accurate localization and well-defined lesion boundaries, even for the most challenging skin lesions. We achieved an average Dice coefficient of 91.18% on the ISBI 2016 Skin Lesion Challenge dataset and 90.66% on the PH2 dataset. Our extensive experimental results on two well-established public benchmark datasets demonstrate that our method is more effective than other state-of-the-art methods for skin lesion segmentation.

First Cold Powering Test of REBCO Roebel Wound Coil for the EuCARD2 Future Magnet Development Project

EuCARD-2 is a project partly supported by FP7-European Commission aiming at exploring accelerator magnet technology for 20 T dipole operating field. The EuCARD-2 collaboration is liaising with similar programs for high field magnets in the USA and Japan. EuCARD-2 focuses, through the work-package 10 “Future magnets,” on the development of a 10 kA-class superconducting, high current density cable suitable for accelerator magnets, for a 5 T stand-alone dipole of 40 mm bore and about 1 m length. After standalone testing, the magnet will possibly be inserted in a large bore background dipole, to be tested at a peak field up to 18 T. This paper starts by reporting on a few of the highlight simulations that demonstrate the progress made in predicting: dynamic current distribution and influence on field quality, complex quench propagation between tapes, and minimum quench energy in the multitape cable. The multiphysics output importantly helps predicting quench signals and guides the development of the novel early detection systems. Knowing current position within individual tapes of each cable we present stress distribution throughout the coils. We report on the development of the mechanical component and assembly processes selected for Feather-M2 the 5 T EuCARD2 magnet. We describe the CERN variable temperature flowing helium cold gas test system. We describe the parallel integration of the FPGA early quench detection system, using pickup coils and temperature sensors, alongside the standard CERN magnet quench detection system using voltage taps. Finally we report on the first cold tests of the REBCO 10 kA class Roebel subscale coil named Feather-M0.