Parallel Devices Research Articles

A Flexible Architecture for Modular Arithmetic Hardware Accelerators based on RNS

Modular arithmetic is a building block for a variety of applications potentially supported on embedded systems. An approach to turn modular arithmetic more efficient is to identify algorithmic modifications that would enhance the parallelization of the target arithmetic in order to exploit the properties of parallel devices and platforms. The Residue Number System (RNS) introduces data-level parallelism, enabling the parallelization even for algorithms based on modular arithmetic with several data dependencies. However, the mapping of generic algorithms to full RNS-based implementations can be complex and the utilization of suitable hardware architectures that are scalable and adaptable to different demands is required. This paper proposes and discusses an architecture with scalability features for the parallel implementation of algorithms relying on modular arithmetic fully supported by the Residue Number System (RNS). The systematic mapping of a generic modular arithmetic algorithm to the architecture is presented. It can be applied as a high level synthesis step for an Application Specific Integrated Circuit (ASIC) or Field Programmable Gate Array (FPGA) design flow targeting modular arithmetic algorithms. An implementation with the Xilinx Virtex 4 and Altera Stratix II Field Programmable Gate Array (FPGA) technologies of the modular exponentiation and Elliptic Curve (EC) point multiplication, used in the Rivest-Shamir-Adleman (RSA) and (EC) cryptographic algorithms, suggests latency results in the same order of magnitude of the fastest hardware implementations of these operations known to date.

Geometric active curve for selective entropy optimization

Recently, with the development of high dimensional large-scale medical imaging devices, the need of fast, robust and accurate segmentation methods is increasing. In this paper, we propose a new level set method (LSM) for image segmentation. The basic idea is to design a selective entropy-based energy functional which is effective and robust against noise, from which we will derive the level set equations and a new selective entropy external forces for the lattice Boltzmann D2Q5 partial differential equation (PDE) solver. The method is accurate and highly parallelizable. The local nature of the lattice Boltzmann method (LBM) allows it to be suitable for fast segmentation methods implemented using some parallel devices such as the graphics processing unit. The proposed algorithm is effective, robust against noise and highly parallelizable. Furthermore, the method can easily be extended to perform an effective image filtering based on Gaussian fuzzy selection. Experimental results on medical images demonstrate subjectively and objectively the performance of the proposed method.

Near-Threshold Computing and Minimum Supply Voltage of Single-Rail MCML Circuits

In high-speed applications, MOS current mode logic (MCML) is a good alternative. Scaling down supply voltage of the MCML circuits can achieve low power-delay product (PDP). However, the current almost all MCML circuits are realized with dual-rail scheme, where the NMOS configuration in series limits the minimum supply voltage. In this paper, single-rail MCML (SRMCML) circuits are described, which can avoid the devices configuration in series, since their logic evaluation block can be realized by only using MOS devices in parallel. The relationship between the minimum supply voltage of the SRMCML circuits and the model parameters of MOS transistors is derived, so that the minimum supply voltage can be estimated before circuit designs. An MCML dynamic flop-flop based on SRMCML is also proposed. The optimization algorithm for near-threshold sequential circuits is presented. A near-threshold SRMCML mode-10 counter based on the optimization algorithm is verified. Scaling down the supply voltage of the SRMCML circuits is also investigated. The power dissipation, delay, and power-delay products of these circuits are carried out. The results show that the near-threshold SRMCML circuits can obtain low delay and small power-delay product.

Рolynomial component PDС-algorithm tasks solving the scheduling theory

This article discusses the usage of the operating shop scheduling in the production management. Some results of the investigation in the sphere are presented in the article. The problem of shop scheduling fulfillment over general due date with identical parallel devices is considered on purpose to maximize a starting moment of devices on condition that all tasks have not been detached. The main goal of the investigation is to determine the latest starting moment of tasks fulfillment with parallel devices over due date in permissible schedule. The article observes the task properties of shop scheduling of tasks fulfillment over general due date with identical parallel devices. According to the method of construction for difficult problems of combinatorial optimization polynomial component is developed PDС-algorithm structure tasks solving. It has been proposed algorithm of initial schedule structure. It has been determined the set of transpositions that allows to improve in succession the significance of the criterion. This set transpositions form the basis of the developed algorithm tasks solving. The results of the investigation could be put into practice of operative shop scheduling of production.

Substantial power density from a discrete nano-scalable biofuel cell

A direct electron transfer biofuel cell consisting of a pair of single walled carbon nanotubes and enzymes on a single plane is reported, which allows for extremely close spacing of electrodes. The discrete device has a measured power and current density of 18mW/cm2 and 90mA/cm2, respectively, into 200TΩ. Analysis of measurements from up to 3million devices in parallel shows that the performance will improve by at least 2 orders of magnitude at peak power.

Fast deterministic switching in orthogonal spin torque devices via the control of the relative spin polarizations

We model 100 ps pulse switching dynamics of orthogonal spin transfer (OST) devices that employ an out-of-plane polarizer and an in-plane polarizer. Simulation results indicate that increasing the spin polarization ratio, CP = PIPP/POPP, results in deterministic switching of the free layer without over-rotation (360° rotation). By using spin torque asymmetry to realize an enhanced effective PIPP, we experimentally demonstrate this behavior in OST devices in parallel to anti-parallel switching. Modeling predicts that decreasing the effective demagnetization field can substantially reduce the minimum CP required to attain deterministic switching, while retaining low critical switching current, Ip ∼ 500 μA.

Neutral Point Clamped Inverter Based PDP Driver

A cost-effective and high-efficiency driver is proposed for plasma display panels (PDPs) by eliminating switching devices with high current ratings. The proposed driver is based on a neutral -point clamped inverter structure, and the path switches can be eliminated without reducing the flexibility in designing driving waveforms. Moreover, the LC resonant operation during the sustain period can be constructed with only unidirectional diodes in aid of the switching scan IC devices. Thus, the proposed driver can produce a PDP driving waveform with no path switches as well as no resonant switches. Since the path switches and the resonant switches are practically implemented with several semiconductor devices in parallel due to high gas-discharge and capacitor displacement current, the proposed driver can considerably simplify the overall driver with fewer semiconductor devices. Additionally, high conduction losses from the path switches and the resonant switches with high current loading can be eliminated in the proposed driver, leading to higher efficiency. The developed driver also places the clamping diodes to recover the inductive energy trapped in the resonant inductors due to the inevitable diode reverse-recovery phenomena.

Nature-Inspired Meta-Heuristics on Modern GPUs: State of the Art and Brief Survey of Selected Algorithms

Graphic processing units (GPUs) emerged recently as an exciting new hardware environment for a truly parallel implementation and execution of Nature and Bio-inspired Algorithms with excellent price-to-power ratio. In contrast to common multicore CPUs that contain up to tens of independent cores, the GPUs represent a massively parallel single-instruction multiple-data devices that can nowadays reach peak performance of hundreds and thousands of giga floating-point operations per second. Nature and Bio-inspired Algorithms implement parallel optimization strategies in which a single candidate solution, a group of candidate solutions (population), or multiple populations seek for optimal solution or set of solutions of given problem. Genetic algorithms (GA) constitute a family of traditional and very well-known nature-inspired populational meta-heuristic algorithms that have proved its usefulness on a plethora of tasks through the years. Differential evolution (DE) is another efficient populational meta-heuristic algorithm for real-parameter optimization. Particle swarm optimization (PSO) can be seen as nature-inspired multiagent method in which the interaction of simple independent agents yields intelligent collective behavior. Simulated annealing (SA) is global optimization algorithm which combines statistical mechanics and combinatorial optimization with inspiration in metallurgy. This survey provides a brief overview of the latest state-of-the-art research on the design, implementation, and applications of parallel GA, DE, PSO, and SA-based methods on the GPUs.

2D and 3D homogenization of laminated cores in the frequency domain

This paper presents 2D and 3D homogenization techniques for laminated iron cores in the frequency domain. In 2D, two approaches (analytical and numerical) are compared in terms of complex reluctivity considering tangential magnetic field to laminations. The analytical approach is generalized to 3D problems. The validation performed on homogenized block gives satisfactory results only in case of closed-core or thin air gap open-core. Thus, we have a correct approximation only when the magnetic field is essentially parallel to laminations. In order to remediate the poor approximation of losses due to non parallel magnetic field component and model large air gap devices, we propose an alternative to block homogenization. A validation is done on a laminated open-core inductance.

Design and analysis of an SOI MEMS voltage step-up converter

This paper presents a comprehensive analysis of a MEMS voltage step-up converter for energy harvesting and other low-power applications. The step-up operation is based on isolating the charge of a mechanically variable capacitor and varying the gap between the electrodes by an appropriate method of providing an actuation force. A bi-stable device is discussed and was specifically designed for static energy harvesting. This device features a separate electrostatic actuator element to manipulate the variable capacitor electrodes. Prototypes were then fabricated using a dicing-free silicon-on-insulator process. The devices have been arbitrarily designed to produce an output voltage which is five times the magnitude of the input (M = 5). Due to leakage currents, it was necessary to cascade the MEMS capacitors in parallel to obtain a higher capacitance level. Parasitic fringing capacitances have a substantial impact on the overall capacitance value of the MEMS device and so the measured multiplication level of the devices is limited to M = 2.125. With four devices in parallel, a maximum output voltage of 35.4 V was obtained for a 24 V input was measured. However, a maximum output voltage of ≈60 V is achievable if the capacitance value was further increased by connecting more devices in parallel or if or the load resistance was increased beyond 1 GΩ.

Low shear stress up-regulation of proinflammatory gene expression in human retinal microvascular endothelial cells

The vascular endothelium responds to shear stress generated by blood flow and changes functions to regulate blood flow and maintain tissue homeostasis. Recently, we found that arteriolar high shear stress leads to increased expression of vasodilatory and antithrombotic genes in human retinal microvascular endothelial cells (HRMECs). However, it is unknown whether low shear stress, which is induced by hypoperfusion particularly in the retinal venules where leukocyte–endothelial interactions mainly occur, affects the retinal endothelial function. We studied the effect of low shear stress on proinflammatory gene expression in HRMECs. The cells were cultured on glass plates and exposed to laminar shear stresses of 0 (static), 1.5 (relatively low flow), and 15 dyne/cm2 (relatively high flow) for 24 h using parallel plate-type flow-loading devices. The mRNA expressions of adhesion molecules, cytokines and chemokines, and procoagulant factors were evaluated using real-time reverse-transcription polymerase chain reaction. HRMECs exposed to 1.5 dyne/cm2 significantly up-regulated the mRNA expression of intercellular adhesion molecule-1, vascular cell adhesion molecule-1, and E-selectin. The cells exposed to 1.5 dyne/cm2 of stress also had increased cytokine/chemokine mRNA expression, i.e., interleukin (IL)-6, IL-8, platelet-derived growth factor-B, and monocyte chemotactic protein-1. Procoagulant factors, i.e., tissue factor and plasminogen activator inhibitor-1 mRNA, increased significantly with exposure to 1.5 dyne/cm2 of stress. Our results showed that relatively low shear stress causes up-regulation of proinflammatory genes in HRMECs, suggesting that decreased shear stress due to vascular hypoperfusion might change the phenotypic characterization of the retinal vascular endothelium and be associated with leukocyte–endothelial interactions.

A novel ethanol gas sensor based on ZnO-microwire

One-dimensional (1D) ZnO microwires were successfully synthesized by chemical vapor deposition and their structural and morphological properties were analyzed by X-ray diffraction and scanning electron microscopy, demonstrating that the microwires were single crystalline with perfect hexagonal structure and smooth surface. Using these 1D microstructures, we fabricated a novel ZnO-based ethanol gas sensor. Operating at room temperature, the sensor was found to have good sensing characteristics. The reliability and stability of the sensor could be improved by connecting multiple 1-wire devices (1-WD) in parallel into a multi-wires device. In interior natural lighting environment and under 3 V bias, the response and recovery time of the 1-WD to 200 ppm ethanol gas were <10 s and about 300 s, respectively, and the minimum and maximum detection limit were about 2 and 200 ppm, respectively. A sensing model was proposed for discussing the performance of the sensor. The simplicity in fabrication, low power consumption and low cost make the sensor suitable for practical application in many fields, especially in identifying driving under the influence and chemical industry monitoring.

Live long and prosper: potentials of low-cost consumer devices for the prevention of cardiovascular diseases.

BackgroundCardiovascular diseases (CVD) are one of the major causes of death worldwide. Personal behavior such as physical activity considerably influences the risk of incurring a CVD. In the last years numerous products such as pedometers have become available on the mass market that allow monitoring relevant behaviors and vital parameters. These devices are sufficiently precise, affordable, and easy to use. While today they are mostly lifestyle oriented they also have considerable potential for health and prevention.ObjectiveOur goal is to investigate how recent low-cost devices can be used in real-life settings for the prevention of CVD, and whether using these devices has an advantage over subjective self-assessment. We also examine whether it is feasible to use multiple of such devices in parallel.MethodsWe observe whether and how persons are willing and able to use multiple devices in their daily lives. We compare the devices’ measurements with subjective self-assessment. We make use of existing low-cost consumer devices to monitor a user's behavior. By mapping the devices' features with pre-defined prevention goals we ensure that the system collects meaningful data that can be used to monitor the individual's behavior. We conducted a user study with 10 healthy adults to measure usability and to identify problems with sensor use in real life. The participants used the devices' original portals to monitor their behavior. The subjects (age range 35-75) used an off-the-shelf pedometer and a sports watch for 4 weeks.ResultsThe participants responded in principle positively to the use of the devices. Analyzing the sensor data, we found that the users had some difficulties in operating the devices. We also found that the participants' self-assessment of their health behavior was too optimistic compared to the monitored data. They rated the usability of the overall system with 71 of up to 100 points in the "System Usability Scale".ConclusionsOur study indicates that today's devices are suitable for a long term monitoring of health for the prevention of CVD. Using the devices provides more precise data than a subjective self-assessment. However usability and acceptance of the systems are still major topics.

GPU-Mapping: Robotic Map Building with Graphical Multiprocessors

This article provides a broad perspective of the potential applicability of graphical processing units (GPUs) computing power in robotics, specifically in the well-known problem of two-dimensional (2-D) robotic mapping. There are three possible ways of exploiting these massively parallel devices: 1) parallelizing existing algorithms, 2) integrating already existing parallelized general purpose software, and 3) use of its high-computational capabilities in the inception of new algorithms. This article presents examples for all three options: parallelizing a popular implementation of the gridmapping algorithm, using a GPU open-source linear sparse system solver to address the problem of linear least squares graph minimization, and developing a novel method that can be efficiently parallelized and executed in a GPU for handling overlapping grid maps in a mapping with local maps algorithm. Large speedups are shown in the experiments, highlighting the importance of this technology in robotic software development in the near future, as is already the case in many other areas.

Multifunctional Graphene–PEDOT Microelectrodes for On‐Chip Manipulation of Human Mesenchymal Stem Cells

AbstractAll‐solution‐processed multifunctional organic bioelectronics composed of reduced graphene oxide (rGO) and dexamethasone 21‐phosphate disodium salt (DEX)‐loaded poly(3,4‐ethylenedioxythiophene) (PEDOT) microelectrode arrays on indium tin oxide glass are reported. They can be used to manipulate the differentiation of human mesenchymal stem cells (hMSCs). In the devices, the rGO material functions as an adhesive coating to promote the adhesion and alignment of hMSC cells and to accelerate their osteogenic differentiation. The poly(L‐lysine‐graft‐ethylene glycol) (PLL‐g‐PEG)‐coated PEDOT electrodes serve as electroactive drug‐releasing electrodes. In addition, the corresponding three‐zone parallel devices operate as efficient drug‐releasing components through spatial‐temporal control of the release of the drug DEX from the PEDOT matrix. Such devices can be used for long‐term cell culturing and controlled differentiation of hMSCs through electrical stimulation.

Solving 3D anisotropic elastic wave equations on parallel GPU devices

Efficiently modeling seismic data sets in complex 3D anisotropic media by solving the 3D elastic wave equation is an important challenge in computational geophysics. Using a stress-stiffness formulation on a regular grid, we tested a 3D finite-difference time-domain solver using a second-order temporal and eighth-order spatial accuracy stencil that leverages the massively parallel architecture of graphics processing units (GPUs) to accelerate the computation of key kernels. The relatively small memory of an individual GPU limits the model domain sizes that can be computed on a single device. To circumvent this constraint and move toward modeling industry-sized 3D anisotropic elastic data sets, we parallelized computation across multiple GPU devices by using domain decomposition and, for each time step, employing an interdevice communication protocol to exchange data values falling within interior boundaries of each subdomain. For two or more GPU devices within a single compute node, we use direct peer-to-peer (i.e., GPU-to-GPU) communication, whereas for networked nodes we employed message-passing interface directives to route data over the network. Our 2D GPU-based anisotropic elastic modeling tests achieved a [Formula: see text] speedup relative to an OpenMP CPU implementation run on an eight-core machine, whereas our 3D tests using dual-GPU devices produced up to a [Formula: see text] speedup. The performance boost afforded by the GPU architecture allowed us to model seismic data for 3D anisotropic elastic models at lower hardware cost and in less time than has been previously possible.

Optimizing data structures in high-level programs

High level data structures are a cornerstone of modern programming and at the same time stand in the way of compiler optimizations. In order to reason about user- or library-defined data structures compilers need to be extensible. Common mechanisms to extend compilers fall into two categories. Frontend macros, staging or partial evaluation systems can be used to programmatically remove abstraction and specialize programs before they enter the compiler. Alternatively, some compilers allow extending the internal workings by adding new transformation passes at different points in the compile chain or adding new intermediate representation (IR) types. None of these mechanisms alone is sufficient to handle the challenges posed by high level data structures. This paper shows a novel way to combine them to yield benefits that are greater than the sum of the parts. Instead of using staging merely as a front end, we implement internal compiler passes using staging as well. These internal passes delegate back to program execution to construct the transformed IR. Staging is known to simplify program generation, and in the same way it can simplify program transformation. Defining a transformation as a staged IR interpreter is simpler than implementing a low-level IR to IR transformer. With custom IR nodes, many optimizations that are expressed as rewritings from IR nodes to staged program fragments can be combined into a single pass, mitigating phase ordering problems. Speculative rewriting can preserve optimistic assumptions around loops. We demonstrate several powerful program optimizations using this architecture that are particularly geared towards data structures: a novel loop fusion and deforestation algorithm, array of struct to struct of array conversion, object flattening and code generation for heterogeneous parallel devices. We validate our approach using several non trivial case studies that exhibit order of magnitude speedups in experiments.

Handling RF Power: The Latest Advances in RF-MEMS Tunable Filters

RF-MEMS technology permits practical implementations of high-Q (Q from 300 to over 1,000) tunable microwave filters with power handling in the >;1-10 W range, with significant tuning range (>;50% to 100%), and ultra-low power consumption. Continuously frequency-tunable filters with mechanical actuation in the opposite direction of RF-inducted pressure can reach powers up to several watts, using a variable compensation bias. Very high Q, and large tuning range can be achieved with this technology, both in planar and cavity tunable filters. Using several devices in parallel, RF-MEMS capacitor arrays can handle RF-powers better than 1 watt, without the need for any compensation tuning.

Acoustic streaming in the transducer plane in ultrasonic particle manipulation devices

In acoustofluidic manipulation and sorting devices, Rayleigh streaming flows are typically found in addition to the acoustic radiation forces. However, experimental work from various groups has described acoustic streaming that occurs in planar devices in a plane parallel to the transducer face. This is typically a four-quadrant streaming pattern with the circulation parallel to the transducer. Understanding its origins is essential for creating designs that limit or control this phenomenon. The cause of this kind of streaming pattern has not been previously explained as it is different from the well-known classical streaming patterns such as Rayleigh streaming and Eckart streaming, whose circulation planes are generally perpendicular to the face of the acoustic transducer. In order to gain insight into these patterns we present a numerical method based on Nyborg's limiting velocity boundary condition that includes terms ignored in the Rayleigh analysis, and verify its predictions against experimental PIV results in a simple device. The results show that the modelled particle trajectories match those found experimentally. Analysis of the dominant terms in the driving equations shows that the origin of this kind of streaming pattern is related to the circulation of the acoustic intensity.

高重复频率脉冲功率技术及其应用:(4)半导体开关的特长与局限性

Power semiconductor devices have been used for repetitive pulsed power generation, due to their advantages in stability, reliability, and lifetime. On the other hand, their limitations in power capability and noise vulnerability are expected to be overcome by using various techniques in gate control and circuit configuration. In this paper, some pulse circuit methods to improve the working ability of semiconductor switches are introduced, including using a number of devices in series and parallel, voltage superposition, combination and complementation of switching devices.