High Throughput Efficiency Research Articles

Regions-of-interest discovering and predicting in smartphone environments

Location-Based Services (LBSs) provide users especially in smartphones with information and services based on their location and interests. Predicting people’s next location could help in improving the quality of such services and, in turn, boosting people’s confidence on those services. Any successful LBS algorithm or model should target three major goals: Location prediction accuracy, high throughput or fast response and efficiency in terms of utilizing smartphone resources. This paper proposes a new approach to discover and predict people’s next location based on their mobility patterns, while being computationally efficient. The approach starts by discovering Regions-of-Interest (RoIs) in people’s historical trajectories (which denote the locations where users have been previously, frequently). A new model based on Markov Chain (MC) is proposed to overcome the drawback of classical MC. Our proposed model considers both space and time contexts where the space represents a specific location that has been visited by a user while the time represents the location visiting time. We show how classical MC model can be extended to include movement times and how time will improve prediction accuracy. One unique finding in our research is related to the value of integrating users’ mobility location/space with time context. In particular, time context is formulated in a way to add extra information to the space context. For better abstraction during building the model, a general transformation function is used to transform the n-order MC into first order. Results show that the proposed approach provides significant improvements in predicting people’s next location compared with the state-of-the-art models when applied on two large real-life datasets, GeoLife and Gowalla.

Tunable particle separation in a hybrid dielectrophoresis (DEP)- inertial microfluidic device

Particle separation is indispensable in many microfluidic systems and holds a broad range of biomedical applications. Inertial microfluidic devices that work solely on intrinsic hydrodynamic forces and inertial effects can offer label-free, high throughput and high efficiency separation performance. However, the working range of the current inertial microfluidic systems is obtained by tailoring the inertial lift forces and secondary flow drag through flow speed. Each channel design is normally effective for specific target particles, which inevitably lacks the flexibility for various particle mixtures. Redesigning the structure and dimension of microchannels for new sets of particle mixtures is often time-consuming and expensive. In this work, by introducing an external dielectrophoretic force field and coupling it with inertial forces, we proposed here an innovative hybrid DEP-inertial microfluidic platform for particle tunable separation. The working principle of the device was explained and its functionality was validated by experiments. In addition, the dimension of target particle mixture can be varied by adjusting the electrical voltage without redesigning the channel structure or dimensions. It is expected that the proposed DEP-inertial concept can work as a flexible platform for a wide range of biomedical applications.

Silver decahedral nanoparticles empowered SPR imaging-SELEX for high throughput screening of aptamers with real-time assessment

A highly efficient method for aptamer screening with real-time monitoring of the SELEX process was described by silver decahedra nanoparticles (Ag10-NPs) enhanced surface plasmon resonance imaging (SPRI). A microarray chip was developed by immobilization of target protein (Lactoferrin (Lac)) and control proteins (α-lactalbumin (α), β-lactoglobulin (β), casein, and bovine serum albumin (BSA)) on the biochip surface. Ag10-NPs were conjugated with an ssDNA library (lib) (Ag10-NPs-library) that consisted of a central 40 nt randomized sequence and a 20 nt fixed primer sequence. Introduction of the Ag10-NPs-library to the SPRI flow channels drastically increased the sensitivity of SPRI signal for real-time monitoring of SELEX. The work allows rapid screening of potential targets, and yields nine aptamers with high affinity (nanomolar range) for Lac after only six-rounds of selection. The aptamer Lac 13-26 was then further tested by SPRI, and the results demonstrated that the aptamer had the capacity to be ultra-sensitive for specific detection of Lac. The novel SPRI-SELEX method demonstrated here showed many advantages of real-time evaluation, high throughput, and high efficiency.

Compact and High Speed Architectures of KASUMI Block Cipher

This paper proposes highly compact and high speed hardware architectures of 64-bit KASUMI block cipher for wide range of wireless applications. A novel methodology is adopted for low area KASUMI implementation employing a single combined substitution block (CSB) for S9 and S7 substitution functions in FI/FO function. The CSB performs S9/S7 transformations and re-utilizes the common combinational logic of AND and AND–XOR gates whereas computations of CSB logic expressions are performed using parallel hierarchy of ANDs/AND-XORs. This scheme substantially reduces the area generating moderate throughput values for compact KASUMI architecture. FPGA implementation with Xilinx Virtex 7 and ASIC implementation using IC Design Compiler, 0.18 µm at 1.8 V constituted 126 CLB slices/2.64 k gates having throughput values of 180/122 Mbps respectively. For high speed KASUMI implementation, the optimized combinational logic methdodology of compact architecture is extended to individual S9/S7 substitution functions and 2 × pipeline schemes are proposed. The odd and even round functions FO and FL are configured for simultaneous operations and path delays are reduced using XOR logic modifications thereby producing high throughput and high efficiency values. Hardware implementations yielded an area of 1619 CLB slices/52.7 k gates attaining throughput values of 16.9/16.1 Gbps for high speed KASUMI architecture I whereas KASUMI architecture II constituted 1847 CLB slices/59.8 k gates attaining very high throughput of 27.5/26.6 Gbps with FPGA and ASICs respectively. A detailed design and performance analysis of proposed KASUMI architectures is described.

Biomonitoring application in the exposome assessment

Biomonitoring can be applied to assess internal exposure and environmental exposure by exposure markers with providing internal exposure to biological characterization and individual exposure information, which is a key tool to evaluate the risk exposure to disease by biological alternation information. With the development of high throughput, broad spectrum and high efficiency screening and detection technology, biomonitoring is defined as traditional biological monitoring (targeted monitoring) and non targeted monitoring analysis (exposomic approaches). An exposomic approach differs from traditional biomonitoring in that it can theoretically include all exposures of potential health significance, whether they are derived from exogenous sources. Both traditional and nontraditional biomonitoring methods should be used to understand the complexity of exposures faced throughout the lifespan. Through hybrid approaches, emerging techniques and the integration of bioinformatics, and developing the detection methods for low abundance chemicals, improving the differentiation ability between endogenous and exogenous chemical, the health outcomes and exposures can be widely recognized and characterized, which can finally contribute to improving the precise prevention and intervention for diseases under the new exposomic model.

Cognitive radio network based DSTTD-CDMA system

In this contribution, the performance of double space time transmit diversity (DSTTD) based multiple-input and multiple-output system with spatial diversity is investigated for cognitive radio network (CRN) based code-division-multiple-access (CDMA) system. CRN is realized in order to invoke available idle spectrum for reliable communication between base station and mobile station. DSTTD is constructed using two space- time processing unit. DSTTD is proven to provide high throughput .The user-specific spreading code is generated using sub-band frequency of CRN. Time-Frequency domain signature sequence is exploited to distinguish secondary users. The performance of CRN based DSTTD-CDMA scheme is analyzed using coded system. At each MS, iterative type of decoder is implemented to estimate secondary users’ information. The performance is analyzed for coded CRN based DSTTD-CDMA system for channel model with Stanford University Interim along with long-term evolution specifications. It is observed through simulation results that coded CR DSTTD-CDMA system establish proficient communication for secondary users while offering higher throughput and higher spectral efficiency and better performance in the context of bit-error-rate with fewer value of signal-to-noise ratio. Further, it is discerned that decoding algorithm obviate adjacent-channel interference along with secondary multi-user interference and CRN based CDMA system in conjunction DSTTD offers solution for high data rate application.

Memory-Efficient Probabilistic 2-D Finite Impulse Response (FIR) Filter

High memory/storage complexity poses severe challenges to achieving high throughput and high energy efficiency in discrete 2-D FIR filtering. This performance bottleneck is especially acute for embedded image or video applications, that use 2-D FIR processing extensively, because real-time processing and low power consumption are their paramount design objectives. Fortunately, most of such perception-based embedded applications possess so-called “inherent fault tolerance”, meaning slight computing accuracy degradation has a little negative effect on their quality of results, but has significant implication to their throughput, hardware implementation cost, and energy efficiency. This paper develops a novel stochastic-based 2-D FIR filtering architecture that exploits the well-known probabilistic convolution theorem to achieve both low hardware cost and high energy efficiency while achieving very high throughput and computing robustness. Our ASIC synthesis results show that stochastic-based architecture achieves $L$ outputs per cycle with 97 and 81 percent less area-delay-product (ADP), and 77 and 67 percent less power consumption compared with the conventional structure and recently published state-of-the-art architecture, respectively, when the 2-D FIR filter size is $4 \times 4$ , the input block size is $L=4$ , and the image size is $512 \times 512$ .

High pressure inertial focusing for separation and concentration of bacteria at high throughput

Inertial focusing is a phenomenon where particles migrate across streamlines in microchannels and focus at well-defined, size dependent equilibrium points of the cross section. It can be taken into advantage for focusing, separation and concentration of particles at high through-put and high efficiency. As particles decrease in size, smaller channels and higher pressures are needed. Hence, new designs are needed to decrease the pressure drop. In this work a novel design was adapted to focus and separate 1 µm from 3 µm spherical polystyrene particles. Also 0.5 µm spherical polystyrene particles were separated, although in a band instead of a single line. The ability to separate, concentrate and focus bacteria, its simplicity of use and high throughput make this technology a candidate for daily routines in laboratories and hospitals.

위성 방송 시스템에서 최적의 고전송 효율 기법 연구

차세대 위성 방송 시스템에서는 빠른 전송 효율과 신뢰도 있는 통신에 대한 요구가 지속적으로 증가하고 있다. 전송 효율을 높이고 성능을 향상시키기 위해 새로운 위성 표준 인 DVB-S3 (Digital Video Broadcasting - Satellite - Third Generation) 시스템에서는 Nyquist 전송률 보다 빠르게 전송하는 FTN (Faster Than Nyquist) 방식이 대두되고 있다. 이러한 요구에 맞춰 본 논문에서는 전송 효율 향상을 위한 세 가지 방식에 대한 성능을 비교 분석하였다. 기존의 LDPC 부호기에서 고 부호화율 방식과 LDPC 복호화 과정에서 생선된 check node를 이용하여 bit node를 puncturing시켜 부호화율 증가 시키는 punctured LDPC 방식, 그리고 Nyquist 속도보다 빠르게 전송하는 FTN 방식이 고려된다. 시뮬레이션을 통해, 세 가지 방식을 동일 전송률 관점에서 성능을 비교한 결과 FTN 방식은 동일한 전송 효율을 유지하면서 성능 면에서 가장 효율적임을 확인하였다. In next generation satellite broadcasting system, requirement of high throughput efficiency has been increasing continuously. To increase throughput efficiency and improve bit error performance simultaneously, FTN method and LDPC codes are employed in new sattelite standard, DVB-S3 system. This paper considered three kinds of methods for increase throughput efficiency. Firstly, as conventional one, high coding rate parity matrix in LDPC encoder is considered. Secondly, punctured coding scheme which delete the coded symbols according to appropriate rules is considered. Lastly, FTN method which transmit fater than Nyquist rate is considered. Among of three kinds of methods, FTN method is most efficient in aspect to performance while maintain same throughput efficiency.

A Review of MAC Scheduling Algorithms in LTE System

The recent wireless communication networks rely on the new technology named Long Term Evolution (LTE) to offer high data rate real-time (RT) traffic with better Quality of Service (QoS) for the increasing demand of customer requirement. LTE provide low latency for real-time services with high throughput, with the help of two-level packet retransmission. Hybrid Automatic Repeat Request (HARQ) retransmission at the Medium Access Control (MAC) layer of LTE networks achieves error-free data transmission. The performance of the LTE networks mainly depends on how effectively this HARQ adopted in the latest communication standard, Universal Mobile Telecommunication System (UMTS). The major challenge in LTE is to balance QoS and fairness among the users. Hence, it is very essential to design a down link scheduling scheme to get the expected service quality to the customers and to utilize the system resources efficiently. This paper provides a comprehensive literature review of LTE MAC layer and six types of QoS/Channel-aware downlink scheduling algorithms designed for this purpose. The contributions of this paper are to identify the gap of knowledge in the downlink scheduling procedure and to point out the future research direction. Based on the comparative study of algorithms taken for the review, this paper is concluded that the EXP Rule scheduler is most suited for LTE networks due to its characteristics of less Packet Loss Ratio (PLR), less Packet Delay (PD), high throughput, fairness and spectral efficiency.

Dynamic configuration management of a multi-standard and multi-mode reconfigurable multi-ASIP architecture for turbo decoding

The multiplication of connected devices goes along with a large variety of applications and traffic types needing diverse requirements. Accompanying this connectivity evolution, the last years have seen considerable evolutions of wireless communication standards in the domain of mobile telephone networks, local/wide wireless area networks, and Digital Video Broadcasting (DVB). In this context, intensive research has been conducted to provide flexible turbo decoder targeting high throughput, multi-mode, multi-standard, and power consumption efficiency. However, flexible turbo decoder implementations have not often considered dynamic reconfiguration issues in this context that requires high speed configuration switching. Starting from this assessment, this paper proposes the first solution that allows frame-by-frame run-time configuration management of a multi-processor turbo decoder without compromising the decoding performances.

Application of a Parallelizable Perfusion Bioreactor for Physiologic 3D Cell Culture

It is crucial but challenging to keep physiologic conditions during the cultivation of 3D cell scaffold constructs for the optimization of 3D cell culture processes. Therefore, we demonstrate the benefits of a recently developed miniaturized perfusion bioreactor together with a specialized incubator system that allows for the cultivation of multiple samples while screening different conditions. Hence, a decellularized bone matrix was tested towards its suitability for 3D osteogenic differentiation under flow perfusion conditions. Subsequently, physiologic shear stress and hydrostatic pressure (HP) conditions were optimized for osteogenic differentiation of human mesenchymal stem cells (MSCs). X-ray computed microtomography and scanning electron microscopy (SEM) revealed a closed cell layer covering the entire matrix. Osteogenic differentiation assessed by alkaline phosphatase activity and SEM was found to be increased in all dynamic conditions. Furthermore, screening of different fluid shear stress (FSS) conditions revealed 1.5 mL/min (equivalent to ∼10 mPa shear stress) to be optimal. However, no distinct effect of HP compared to flow perfusion without HP on osteogenic differentiation was observed. Notably, throughout all experiments, cells cultivated under FSS or HP conditions displayed increased osteogenic differentiation, which underlines the importance of physiologic conditions. In conclusion, the bioreactor system was used for biomaterial testing and to develop and optimize a 3D cell culture process for the osteogenic differentiation of MSCs. Due to its versatility and higher throughput efficiency, we hypothesize that this bioreactor/incubator system will advance the development and optimization of a variety of 3D cell culture processes.

Dean Flow Dynamics in Low-Aspect Ratio Spiral Microchannels

A wide range of microfluidic cell-sorting devices has emerged in recent years, based on both passive and active methods of separation. Curvilinear channel geometries are often used in these systems due to presence of secondary flows, which can provide high throughput and sorting efficiency. Most of these devices are designed on the assumption of two counter rotating Dean vortices present in the curved rectangular channels and existing in the state of steady rotation and amplitude. In this work, we investigate these secondary flows in low aspect ratio spiral rectangular microchannels and define their development with respect to the channel aspect ratio and Dean number. This work is the first to experimentally and numerically investigate Dean flows in microchannels for Re > 100, and show presence of secondary Dean vortices beyond a critical Dean number. We further demonstrate the impact of these multiple vortices on particle and cell focusing. Ultimately, this work offers new insights into secondary flow instabilities for low-aspect ratio, spiral microchannels, with improved flow models for design of more precise and efficient microfluidic devices for applications such as cell sorting and micromixing.

IEEE 802.15.4 Frame Aggregation Enhancement to Provide High Performance in Life-Critical Patient Monitoring Systems.

In wireless body area sensor networks (WBASNs), Quality of Service (QoS) provision for patient monitoring systems in terms of time-critical deadlines, high throughput and energy efficiency is a challenging task. The periodic data from these systems generates a large number of small packets in a short time period which needs an efficient channel access mechanism. The IEEE 802.15.4 standard is recommended for low power devices and widely used for many wireless sensor networks applications. It provides a hybrid channel access mechanism at the Media Access Control (MAC) layer which plays a key role in overall successful transmission in WBASNs. There are many WBASN’s MAC protocols that use this hybrid channel access mechanism in variety of sensor applications. However, these protocols are less efficient for patient monitoring systems where life critical data requires limited delay, high throughput and energy efficient communication simultaneously. To address these issues, this paper proposes a frame aggregation scheme by using the aggregated-MAC protocol data unit (A-MPDU) which works with the IEEE 802.15.4 MAC layer. To implement the scheme accurately, we develop a traffic patterns analysis mechanism to understand the requirements of the sensor nodes in patient monitoring systems, then model the channel access to find the performance gap on the basis of obtained requirements, finally propose the design based on the needs of patient monitoring systems. The mechanism is initially verified using numerical modelling and then simulation is conducted using NS2.29, Castalia 3.2 and OMNeT++. The proposed scheme provides the optimal performance considering the required QoS.

Eyeriss: An Energy-Efficient Reconfigurable Accelerator for Deep Convolutional Neural Networks

Eyeriss is an accelerator for state-of-the-art deep convolutional neural networks (CNNs). It optimizes for the energy efficiency of the entire system, including the accelerator chip and off-chip DRAM, for various CNN shapes by reconfiguring the architecture. CNNs are widely used in modern AI systems but also bring challenges on throughput and energy efficiency to the underlying hardware. This is because its computation requires a large amount of data, creating significant data movement from on-chip and off-chip that is more energy-consuming than computation. Minimizing data movement energy cost for any CNN shape, therefore, is the key to high throughput and energy efficiency. Eyeriss achieves these goals by using a proposed processing dataflow, called row stationary (RS), on a spatial architecture with 168 processing elements. RS dataflow reconfigures the computation mapping of a given shape, which optimizes energy efficiency by maximally reusing data locally to reduce expensive data movement, such as DRAM accesses. Compression and data gating are also applied to further improve energy efficiency. Eyeriss processes the convolutional layers at 35 frames/s and 0.0029 DRAM access/multiply and accumulation (MAC) for AlexNet at 278 mW (batch size $N = 4$ ), and 0.7 frames/s and 0.0035 DRAM access/MAC for VGG-16 at 236 mW ( $N = 3$ ).

Cross-Networks Energy Efficiency Tradeoff: From Wired Networks to Wireless Networks

The tremendously large number of increasing Internet protocol (IP) packets call for quality of service (QoS) guaranteed packets transmission with low-delay, high throughput, and high energy efficiency (defined as the transmitted bits per unit energy consumption) in the fifth-generation (5G) networks. For this motivation, wavelength division multiplexing (WDM) networks and the next generation of wireless technologies are two major methods in the wired and wireless networks, respectively. However, the existing energy efficient switch fabric and wireless technologies have focused on either wired or wireless networks, only separately. The joint cross-networks optimization for energy efficiency in 5G remains unexplored. This does not fully facilitate the QoS guaranteed packets transmission and energy efficient networks planning from the viewpoint of cross networks. In this paper, we formulate a joint optimization model to enhance the performance of energy efficiency in 5G. In particular, each base station is equipped with a set of parallel tunable lasers for simultaneous transmission of multiple packets from the uplink users in the cell as well as the data center networks. We propose a novel joint cross-networks scheduling and routing (JCNSR) algorithm according to the wireless channel quality of users, user data rate, and the topology constraint. The IP packets are then delivered to the targeted cells via a transport layer in the WDM network and further transmitted to the targeted users via wireless channels under the constraint of delay. Based on the idea of the cross-networks tradeoff between the delay and the energy efficiency, JCNSR can achieve high-energy efficient transmission with performance guarantee. The effectiveness of the proposed framework is verified by extensive simulations.

GPUvm: GPU Virtualization at the Hypervisor

Graphic processing units (GPUs) provide a massively-parallel computational power and encourage the use of general-purpose computing on GPUs (GPGPU). The distinguished design of <i>discrete GPUs</i> helps them to provide the high throughput, scalability, and energy efficiency needed for GPGPU applications. Despite the previous study on GPU virtualization, the tradeoffs between the virtualization approaches remain unclear, because of a lack of designs for or quantitative evaluations of the hypervisor-level virtualization for discrete GPUs. Shedding light on these tradeoffs and the technical requirements for the hypervisor-level virtualization would facilitate the development of an appropriate GPU virtualization solution. <italic/> <inline-formula><tex-math notation="LaTeX"> $\sf{GPUvm}$</tex-math> </inline-formula> <italic/> , which is an open architecture for hypervisor-level GPU virtualization with a particular emphasis on using the Xen hypervisor, is presented in this paper. <inline-formula><tex-math notation="LaTeX"> $\sf{GPUvm}$</tex-math> </inline-formula> offers three virtualization modes: the full-, naive para-, and high-performance para-virtualization. <inline-formula><tex-math notation="LaTeX">$\sf{GPUvm}$</tex-math></inline-formula> exposes low- and high-level interfaces such as memory-mapped I/O and DRM APIs to the guest virtual machines (VMs). Our experiments using a relevant commodity GPU showed that <inline-formula><tex-math notation="LaTeX">$\sf{GPUvm}$</tex-math></inline-formula> incurs different overheads as the level of the exposed interfaces is changed. The results also showed that a coarse-grained fairness on the GPU among multiple VMs can be achieved using GPU scheduling.

Enhanced Exfoliation Effect of Solid Auxiliary Agent On the Synthesis of Biofunctionalized MoS2Using Grindstone Chemistry

Mass production and commercial availability are prerequisites for the viability and wide application of MoS2. Here, we demonstrate enhanced grindstone chemistry for a one-step synthesis of biofunctionalized MoS2. By adding a SiO2 auxiliary agent the exfoliation efficiency increases from 16.23% to 58.59% and a rapid and high-yield exfoliation of MoS2 is seen. SiO2 exhibits a fragmentation effect, which reduces the lateral size and facilitates the exfoliation of MoS2, thus inducing a high-efficient paradigm in the top-down fabrication of biofunctionalized MoS2 nanosheets. The as-prepared MoS2-chitosan (MoS2-CS) nanosheets display complete disaggregation and homogeneous dispersion, as well as a high content of chitosan (ca. 20 wt%). As a proof-of-concept application, the MoS2-CS nanosheets act as a biosorbent for PbII removal, exhibiting a good adsorption capacity and recyclability. This green and facile enhanced grindstone chemistry with minimal use of organic solvents and high-throughput efficiency can be extended to the fabrication of other biocompatible inorganic 2D analogues for a variety of applications.

A Novel Flow Regulation Protocol to Optimize the End-to-End Performance and Fairness Over LEO Satellite Network

The major challenge in the utilization of satellite network for the Internet access is its differing nature with wired network. To increase the performance of TCP-based applications over satellite network, performance enhancement proxies (PEP) had been incorporated. However, the existing PEP solutions have violated the end-to-end semantics of Internet applications. To solve this problem and to meet the future needs of the communication network, next-generation transport architecture (Iyengar and Ford in flow splitting with fate sharing in a next-generation transport services architecture, 2009; in A next-generation transport services architecture, 2009. http://tools.ietf.org/html/draft-iyengar-ford-tng-00) has been proposed. In this architecture, flow regulation functions (flow regulation layer) are factored out from end-to-end best delivery (endpoint layer). As a contribution to this research direction, we propose a novel flow regulation protocol to control the flows in terms of congestion, error and fairness over the Low Earth Orbit satellite environment. Simulation results reveal that the proposed protocol can provide optimized end-to-end performance (high throughput, high goodput and high information efficiency) without violating end-to-end semantics, and it also meets the network-level objective of fairness among the flows.

FPGA‐accelerated deep convolutional neural networks for high throughput and energy efficiency

SummaryRecent breakthroughs in the deep convolutional neural networks (CNNs) have led to great improvements in the accuracy of both vision and auditory systems. Characterized by their deep structures and large numbers of parameters, deep CNNs challenge the computational performance of today. Hardware specialization in the form of field‐programmable gate array offers a promising path towards major leaps in computational performance while achieving high‐energy efficiency.In this paper, we focus on accelerating deep CNNs using the Xilinx Zynq‐zq7045 FPGA SoC. As most of the computational workload can be converted to matrix multiplications, we adopt a matrix multiplier‐based accelerator architecture. Dedicated units are designed to eliminate the conversion overhead. We also design a customized memory system according to the memory access pattern of CNNs. To make the accelerator easily usable by application developers, our accelerator supports Caffe, which is a widely used software framework of deep CNN. Different CNN models can be adopted by our accelerator, with good performance portability. The experimental results show that for a typical application of CNN, image classification, an average throughout of 77.8 GFLOPS is achieved, while the energy efficiency is 4.7× better than an Nvidia K20 GPGPU. © 2016 The Authors. Concurrency and Computation: Practice and Experience Published by John Wiley &amp; Sons Ltd