Parallel Processing Algorithm Research Articles

FPGA implementation of image processing technique for blood samples characterization

This work presents a hardware implementation of an image processing algorithm for blood type determination. The image processing technique proposed in this paper uses the appearance of agglutination to determine blood type by detecting edges and contrast within the agglutinated sample. An FPGA implementation and parallel processing algorithms are used in conjugation with image processing techniques to make this system reliable for the characterization of large numbers of blood samples. The program was developed using Matlab software then transferred and implemented on a Vertex 6 FPGA from Xilinx employing ISE software. Hardware implementation of the proposed algorithm on FPGA demonstrates a power consumption of 770mW from a 2.5V power supply. Blood type characterization using our FPGA implementation requires only 6.6s, while a desktop computer-based algorithm with Matlab implementation on a Pentium 4 processor with a 3GHz clock takes 90s. The presented device is faster, more portable, less expensive, and consumes less power than conventional instruments. The proposed hardware solution achieved accuracy of 99.5% when tested with over 500 different blood samples.

MODYLAS: A Highly Parallelized General-Purpose Molecular Dynamics Simulation Program for Large-Scale Systems with Long-Range Forces Calculated by Fast Multipole Method (FMM) and Highly Scalable Fine-Grained New Parallel Processing Algorithms.

Our new molecular dynamics (MD) simulation program, MODYLAS, is a general-purpose program appropriate for very large physical, chemical, and biological systems. It is equipped with most standard MD techniques. Long-range forces are evaluated rigorously by the fast multipole method (FMM) without using the fast Fourier transform (FFT). Several new methods have also been developed for extremely fine-grained parallelism of the MD calculation. The virtually buffering-free methods for communications and arithmetic operations, the minimal communication latency algorithm, and the parallel bucket-relay communication algorithm for the upper-level multipole moments in the FMM realize excellent scalability. The methods for blockwise arithmetic operations avoid data reload, attaining very small cache miss rates. Benchmark tests for MODYLAS using 65 536 nodes of the K-computer showed that the overall calculation time per MD step including communications is as short as about 5 ms for a 10 million-atom system; that is, 35 ns of simulation time can be computed per day. The program enables investigations of large-scale real systems such as viruses, liposomes, assemblies of proteins and micelles, and polymers.

HAPLOFIND: A New Method for High-Throughput mtDNA Haplogroup Assignment

Deep sequencing technologies are completely revolutionizing the approach to DNA analysis. Mitochondrial DNA (mtDNA) studies entered in the "postgenomic era": the burst in sequenced samples observed in nuclear genomics is expected also in mitochondria, a trend that can already be detected checking complete mtDNA sequences database submission rate. Tools for the analysis of these data are available, but they fail in throughput or in easiness of use. We present here a new pipeline based on previous algorithms, inherited from the "nuclear genomic toolbox," combined with a newly developed algorithm capable of efficiently and easily classify new mtDNA sequences according to PhyloTree nomenclature. Detected mutations are also annotated using data collected from publicly available databases. Thanks to the analysis of all freely available sequences with known haplogroup obtained from GenBank, we were able to produce a PhyloTree-based weighted tree, taking into account each haplogroup pattern conservation. The combination of a highly efficient aligner, coupled with our algorithm and massive usage of asynchronous parallel processing, allowed us to build a high-throughput pipeline for the analysis of mtDNA sequences that can be quickly updated to follow the ever-changing nomenclature. HaploFind is freely accessible at the following Web address: https://haplofind.unibo.it.

Parallel Processing for Large-scale Fault Tree in Wireless Sensor Networks

Wireless sensor networks (WSN) covers many kinds of technologies, such as technology of sensor, embedded system, wireless communication, etc. WSN is different from the traditional networks in size, communication distance and energy-constrained so as to develop new topology, protocol, quality of service (QoS), and so on. In order to solve the problem of self-organizing in the topology, this paper proposes a novel strategy which is based on communication delay between sensors. Firstly, the gateway selects some boundary nodes to connect. Secondly, the boundary nodes choose inner nodes. The rest may be deduced by analogy. Finally, a net-tree topology with multi-path routing is developed. The analyses of the topology show that net-tree has strong ability in self-organizing and extensible. However, the scale of system is usually very large and complexity so that it is hard to detect the failure nodes when the nodes fail. To solve the greater challenge, the paper proposes to adopt fault tree analysis. Fault tree is a commonly used method to analyze the reliability of a network or system. Based on the fault tree analysis, a parallel computing algorithm is represented to these faults in the net-tree. Firstly, two models for parallel processing are came up and we focus on the parallel processing algorithm based on the cut sets. Then, the speedup ratio is studied. Compare with the serial algorithm, the results of the experiment shows that the efficiency has been greatly improved.

Compressed sensing parallel processing algorithm based on OpenMP

Concerning the high complexity and long-time running of the compressed sensing reconstructed algorithm,a compressed sensing parallel algorithm based on multi-core processors was proposed.On the basis of a careful analysis of the compressed sensing algorithm,OpenMP was used for compressed sensing measurement and Orthogonal Matching Pursuit(OMP) algorithm for parallel processing to improve program performance.The experimental results show that the speedup is in linear growth with the increasing threads.The execution of the procedure is more effective.Moreover,the more complex the reconstruction process is,the more obvious the performance optimization will be.

Solving Large Nonlinear Systems of First-Order Ordinary Differential Equations With Hierarchical Structure Using Multi-GPGPUs and an Adaptive Runge Kutta ODE Solver

The adaptive Runge–Kutta (ARK) method on multi-general-purpose graphical processing units (GPUs) is used for solving large nonlinear systems of first-order ordinary differential equations (ODEs) with over <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex Notation="TeX">${\sim}{10\, 000}$</tex></formula> variables describing a large genetic network in systems biology for the biological clock. To carry out the computation of the trajectory of the system, a hierarchical structure of the ODEs is exploited, and an ARK solver is implemented in compute unified device architecture/C++ (CUDA/C++) on GPUs. The result is a 75-fold speedup for calculations of 2436 independent modules within the genetic network describing clock function relative to a comparable CPU architecture. These 2436 modules span one-quarter of the entire genome of a model fungal system, Neurospora crassa. The power of a GPU can in principle be harnessed by using warp-level parallelism, instruction level parallelism or both of them. Since the ARK ODE solver is entirely sequential, we propose a new parallel processing algorithm using warp-level parallelism for solving <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">${\sim}{10\,000}$</tex></formula> ODEs that belong to a large genetic network describing clock genome-level dynamics. A video is attached illustrating the general idea of the method on GPUs that can be used to provide new insights into the biological clock through single cell measurements on the clock.

Multistep Scheduling Algorithm for Parallel and Distributed Processing in Heterogeneous Systems with Communication Costs

We discuss a task scheduling problem in heterogeneous systems and propose a multistep scheduling algorithm to solve it. Existing scheduling algorithms formulated as 0-1 integer linear programming can be used to consider the optimality of task scheduling. However, they cannot address complicated relations among tasks or communication costs among processors. Therefore, we propose a scheduling algorithm that formulates communication costs as 0-1 integer linear programming. On the other hand, 0-1 integer linear programming takes a long time to calculate the scheduling results because it is NP-complete. Thus, scheduling time also needs to be decreased. A solution for decreasing scheduling time is a graph clustering which decomposes a large task graph into smaller subtask graphs (clusters). Also, it is important for parallel and distributed processing to find task parallelism in a task graph. Then, we also propose a <i>clustering algorithm based on SCAN</i>. SCAN is an algorithm for finding clusters in a network. The <i>clustering algorithm based on SCAN</i> can find task parallelism in a task graph. In numerical examples, we argue the following two points. First, our multistep scheduling algorithm resolves the scheduling problem in heterogeneous systems. Second, it is superior to the existing scheduling algorithms in terms of calculation time.

Research on the Parallel Processing Algorithm of STAP Based on Fine-grained Task Scheduling

对于空时自适应信号处理(Space-Time Adaptive Processing, STAP)算法的并行处理问题，传统方法以粗粒度的划分方式将STAP算法分配到特定硬件系统中的不同处理器中，利用处理器间的流水计算来提高系统计算吞吐量。该文分析了传统并行处理方法的缺陷：粗粒度的任务划分方式牺牲了STAP算法的并行度；传统处理方法仅能适用于特定的系统环境。针对上述情况，该文提出一种基于细粒度任务分配的STAP并行处理方法，该方法分为以下3个步骤：构建细粒度的DAG(Direct Acyclic Graph)形式的STAP算法任务模型；使用统一拓扑结构模型描述不同结构的目标硬件系统；基于细粒度任务分配算法将任务模型分配到拓扑结构模型中的处理器实现并行计算。实验结果表明该并行处理方法能够达到良好的加速比，并且对于不同的STAP应用系统具有很好的适应性。

Computing effective properties of random heterogeneous materials on heterogeneous parallel processors

In recent decades, finite element (FE) techniques have been extensively used for predicting effective properties of random heterogeneous materials. In the case of very complex microstructures, the choice of numerical methods for the solution of this problem can offer some advantages over classical analytical approaches, and it allows the use of digital images obtained from real material samples (e.g., using computed tomography). On the other hand, having a large number of elements is often necessary for properly describing complex microstructures, ultimately leading to extremely time-consuming computations and high memory requirements. With the final objective of reducing these limitations, we improved an existing freely available FE code for the computation of effective conductivity (electrical and thermal) of microstructure digital models. To allow execution on hardware combining multi-core CPUs and a GPU, we first translated the original algorithm from Fortran to C, and we subdivided it into software components. Then, we enhanced the C version of the algorithm for parallel processing with heterogeneous processors. With the goal of maximizing the obtained performances and limiting resource consumption, we utilized a software architecture based on stream processing, event-driven scheduling, and dynamic load balancing. The parallel processing version of the algorithm has been validated using a simple microstructure consisting of a single sphere located at the centre of a cubic box, yielding consistent results. Finally, the code was used for the calculation of the effective thermal conductivity of a digital model of a real sample (a ceramic foam obtained using X-ray computed tomography). On a computer equipped with dual hexa-core Intel Xeon X5670 processors and an NVIDIA Tesla C2050, the parallel application version features near to linear speed-up progression when using only the CPU cores. It executes more than 20 times faster when additionally using the GPU.

Editorial: enabling technologies for programming extreme scale systems

Multi-core architecture presents a new trend, and core-based parallel processing algorithms will continue to become more important. In addition, Grid and Cloud Computing (GCC) has emerged rapidly as an exciting new paradigm that offers a challenging model of computing and poses fascinating problems ranging from parallel architecture to programming models and compiler optimization. Furthermore, the wide-scale deployment of computing nodes, clusters, and embedded processors will continue to fuel core parallel processing and distributed systems algorithms. Thus, the importance of scalable parallel programming will continue to grow. The field of parallel and distributed computing is going through rapid changes. The emerging multi-core commodity processors and advanced networking technologies are allowing for the design of cost-effective parallel and distributed systems. To achieve scalability and high performance in these systems, it is increasingly becoming challenging to design better architecture, software environments, networking protocols, compilers, operating systems, languages, algorithms and applications for designing next generation scalable and high-performance parallel systems. This special issue is intended to foster stateof-the-art research in the area of extreme scale parallel systems including the topics of parallel algorithm design, data flow analysis, virtualization, energy efficient, load balancing, parallel computing model, multi-core programming, GPU implementation and optimization, execution on real-world parallel architecture and novel applications associated with this new paradigm. The papers included in this special issue demonstrate the effectiveness and efficiency of a variety of technologies and applications in parallel, grid, cloud and extreme scale systems. All of them not only contribute valuable insights and results but

Parallel implementation of all-digital timing recovery for high-speed and real-time optical coherent receivers

The digital coherent receivers combine coherent detection with digital signal processing (DSP) to compensate for transmission impairments, and therefore are a promising candidate for future high-speed optical transmission system. However, the maximum symbol rate supported by such real-time receivers is limited by the processing rate of hardware. In order to cope with this difficulty, the parallel processing algorithms is imperative. In this paper, we propose a novel parallel digital timing recovery loop (PDTRL) based on our previous work. Furthermore, for increasing the dynamic dispersion tolerance range of receivers, we embed a parallel adaptive equalizer in the PDTRL. This parallel joint scheme (PJS) can be used to complete synchronization, equalization and polarization de-multiplexing simultaneously. Finally, we demonstrate that PDTRL and PJS allow the hardware to process 112 Gbit/s POLMUX-DQPSK signal at the hundreds MHz range.

One-to-One Embedding between Honeycomb Mesh and Petersen-Torus Networks

As wireless mobile telecommunication bases organize their structure using a honeycomb-mesh algorithm, there are many studies about parallel processing algorithms like the honeycomb mesh in Wireless Sensor Networks. This paper aims to study the Peterson-Torus graph algorithm in regard to the continuity with honeycomb-mesh algorithm in order to apply the algorithm to sensor networks. Once a new interconnection network is designed, parallel algorithms are developed with huge research costs to use such networks. If the old network is embedded in a newly designed network, a developed algorithm in the old network is reusable in a newly designed network. Petersen-Torus has been designed recently, and the honeycomb mesh has already been designed as a well-known interconnection network. In this paper, we propose a one-to-one embedding algorithm for the honeycomb mesh (HMn) in the Petersen-Torus PT(n,n), and prove that dilation of the algorithm is 5, congestion is 2, and expansion is 5/3. The proposed one-to-one embedding is applied so that processor throughput can be minimized when the honeycomb mesh algorithm runs in the Petersen-Torus.

GPU-based real-time small displacement estimation with ultrasound

General purpose computing on graphics processing units (GPUs) has been previously shown to speed up computationally intensive data processing and image reconstruction algorithms for computed tomography (CT), magnetic resonance (MR), and ultrasound images. Although some algorithms in ultrasound have been converted to GPU processing, many investigative ultrasound research systems still use serial processing on a single CPU. One such ultrasound modality is acoustic radiation force impulse (ARFI) imaging, which investigates the mechanical properties of soft tissue. Traditionally, the raw data are processed offline to estimate the displacement of the tissue after the application of radiation force. It is highly advantageous to process the data in real-time to assess their quality and make modifications during a study. In this paper, we present algorithms for efficient GPU parallel processing of two widely used tools in ultrasound: cubic spline interpolation and Loupas' two-dimensional autocorrelator for displacement estimation. It is shown that a commercially available graphics card can be used for these computations, achieving speed increases up to 40x compared with single CPU processing. Thus, we conclude that the GPU-based data processing approach facilitates real-time (i.e., <1 second) display of ARFI data and is a promising approach for ultrasonic research systems.

Processing and rendering of Fourier domain optical coherence tomography images at a line rate over 524 kHz using a graphics processing unit

In Fourier domain optical coherence tomography (FD-OCT), a large amount of interference data needs to be resampled from the wavelength domain to the wavenumber domain prior to Fourier transformation. We present an approach to optimize this data processing, using a graphics processing unit (GPU) and parallel processing algorithms. We demonstrate an increased processing and rendering rate over that previously reported by using GPU paged memory to render data in the GPU rather than copying back to the CPU. This avoids unnecessary and slow data transfer, enabling a processing and display rate of well over 524,000 A-scan∕s for a single frame. To the best of our knowledge this is the fastest processing demonstrated to date and the first time that FD-OCT processing and rendering has been demonstrated entirely on a GPU.

Correctness of Specialization-Based Parallel Processing Algorithm for Constraint Satisfaction Problem

The purpose of this paper is to discuss the correctness of a method for constructing parallel processing programs from a problem description. The framework we adopt for this purpose is Equivalent Transformation Framework (ETF), which regards computation as transformation of definite clauses. In the framework, a problem's domain knowledge and a query are described in definite clauses, and its meaning is defined by a model of the set of definite clauses. Then meaning-preserving transformation rules for the query are generated. We propose a parallel processing method based on “specialization”, a part of operation in the transformations, and discuss new parallel processing method based on the specialization that maintains correctness of the computation. The specialization is generalized notion of substitution in logic programming, and it allows more rich representation. We examine the correctness of our approach based the specialization.

Improving the scalability of hyperspectral imaging applications on heterogeneous platforms using adaptive run-time data compression

Latest generation remote sensing instruments (called hyperspectral imagers) are now able to generate hundreds of images, corresponding to different wavelength channels, for the same area on the surface of the Earth. In previous work, we have reported that the scalability of parallel processing algorithms dealing with these high-dimensional data volumes is affected by the amount of data to be exchanged through the communication network of the system. However, large messages are common in hyperspectral imaging applications since processing algorithms are pixel-based, and each pixel vector to be exchanged through the communication network is made up of hundreds of spectral values. Thus, decreasing the amount of data to be exchanged could improve the scalability and parallel performance. In this paper, we propose a new framework based on intelligent utilization of wavelet-based data compression techniques for improving the scalability of a standard hyperspectral image processing chain on heterogeneous networks of workstations. This type of parallel platform is quickly becoming a standard in hyperspectral image processing due to the distributed nature of collected hyperspectral data as well as its flexibility and low cost. Our experimental results indicate that adaptive lossy compression can lead to improvements in the scalability of the hyperspectral processing chain without sacrificing analysis accuracy, even at sub-pixel precision levels.

Nonuniformity correction algorithm based on adaptive filter for infrared focal plane arrays

Response nonuniformity is a key problem that influences the imaging performance of infrared focal plane arrays (IRFPA) imaging system. A parallel processing algorithm to adaptively estimate the nonuniformity correction (NUC) parameters for IRFPA is presented. In this algorithm, a bank of the adaptive filter is applied to adaptively estimate the NUC parameters for every detector in IRFPA. The infrared image sequences are input into the bank of adaptive filter. After certain times recursion calculations are executed frame-by-frame, then the optimal coefficients of the gain and the offset of detector in IRFPA are achieved. Then the NUC is fulfilled ultimately. The algorithm reduces the influence that the response drift with time imposed on NUC effectively, and achieves good NUC effect. It was validated by real experimental imaging procedures.

Markov Chain Monte Carlo Detectors for Channels With Intersymbol Interference

In this paper, we propose novel low-complexity soft-in soft-out (SISO) equalizers using the Markov chain Monte Carlo (MCMC) technique. We develop a bitwise MCMC equalizer (b-MCMC) that adopts a Gibbs sampler to update one bit at a time, as well as a group-wise MCMC (g-MCMC) equalizer where multiple symbols are updated simultaneously. The g-MCMC equalizer is shown to outperform both the b-MCMC and the linear minimum mean square error (MMSE) equalizer significantly for channels with severe amplitude distortion. Direct application of MCMC to channel equalization requires sequential processing which leads to long processing delay. We develop a parallel processing algorithm that reduces the processing delay by orders of magnitude. Numerical results show that both the sequential and parallel processing MCMC equalizers perform similarly well and achieve a performance that is only slightly worse than the optimum maximum <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a posteriori</i> (MAP) equalizer. The MAP equalizer, on the other hand, has a complexity that grows exponentially with the size of the memory of the channel, while the complexity of the proposed MCMC equalizers grows linearly.

Multichannel Spatial Encoding and Extraction for Low-Noise Laser Applications

A multichannel spatial encoding and extraction technique is demonstrated for laser-based experimental applications. The technique makes use of a multichannel beam chopper to encode spatial information onto the transverse profile of a laser beam. Lock-in detection is subsequently employed in a parallel processing algorithm to extract the encoded experimental spatial information. The technique is demonstrated with an optical system incorporating ten parallel frequency channels.

General background and approach to multibody dynamics for space applications

Multibody dynamics for space applications is dictated by space environment such as space-varying gravity forces, orbital and attitude perturbations, control forces if any. Several methods and formulations devoted to the modeling of flexible bodies undergoing large overall motions were developed in recent years. Most of these different formulations were aimed to face one of the main problems concerning the analysis of spacecraft dynamics namely the reduction of computer simulation time. By virtue of this, the use of symbolic manipulation, recursive formulation and parallel processing algorithms were proposed. All these approaches fall into two categories, the one based on Newton/Euler methods and the one based on Lagrangian methods; both of them have their advantages and disadvantages although in general, Newtonian approaches lend to a better understanding of the physics of problems and in particular of the magnitude of the reactions and of the corresponding structural stresses. Another important issue which must be addressed carefully in multibody space dynamics is relevant to a correct choice of kinematics variables. In fact, when dealing with flexible multibody system the resulting equations include two different types of state variables, the ones associated with large (rigid) displacements and the ones associated with elastic deformations. These two sets of variables have generally two different time scales if we think of the attitude motion of a satellite whose period of oscillation, due to the gravity gradient effects, is of the same order of magnitude as the orbital period, which is much bigger than the one associated with the structural vibration of the satellite itself. Therefore, the numerical integration of the equations of the system represents a challenging problem. This was the abstract and some of the arguments that Professor Paolo Santini intended to present for the Breakwell Lecture; unfortunately a deadly disease attacked him and shortly took him to death, leaving his work unfinished. In agreement with Astrodynamics Committee it was decided to prepare a paper based on some research activities that Paolo Santini performed during almost 50 years in the aerospace field. His researches spanned many arguments, encompassing flexible space structures, to optimization, stability analysis, thermal analysis, smart structure, etc. just to mention the ones more related to the space field (Paolo Santini was also one the pioneers of the studies of composite wing structures, aeroelasticity and unsteady aerodynamics for aeronautical applications). Following notes have been prepared by Paolo Gasbarri who was one of Paolo Santini collaborators for almost 15 years, they will attempt to offer a sketch of Professor Santini's activity by focusing on three main topics: the stability of flexible spacecrafts, the dynamics of multibody systems and the use of the smart structure technology for the space applications.