Air pollution modelling using a Graphics Processing Unit with CUDA

The Graphics Processing Unit (GPU) is a powerful tool for parallel computing. In the past years the performance and capabilities of GPUs have increased, and the Compute Unified Device Architecture (CUDA) - a parallel computing architecture - has been developed by NVIDIA to utilize this performance in general purpose computations. Here we show for the first time a possible application of GPU for environmental studies serving as a basement for decision making strategies. A stochastic Lagrangian particle model has been developed on CUDA to estimate the transport and the transformation of the radionuclides from a single point source during an accidental release. Our results show that parallel implementation achieves typical acceleration values in the order of 80-120 times compared to CPU using a single-threaded implementation on a 2.33 GHz desktop computer. Only very small differences have been found between the results obtained from GPU and CPU simulations, which are comparable with the effect of stochastic transport phenomena in atmosphere. The relatively high speedup with no additional costs to maintain this parallel architecture could result in a wide usage of GPU for diversified environmental applications in the near future.     

Particle-field decomposition and domain decomposition in parallel particle-in-cell beam dynamics simulation

Particle-in-cell (PIC) simulation is widely used in many branches of physics and engineering. In this paper, we give an analysis of the particle-field decomposition method and the domain decomposition method in parallel particle-in-cell beam dynamics simulation. The parallel performance of the two decomposition methods was studied on the Cray XT4 and the IBM Blue Gene/P Computers. The domain decomposition method shows better scalability but is slower than the particle-field decomposition in most cases (up to a few thousand processors) for macroparticle dominant applications. The particle-field decomposition method also shows less memory usage than the domain decomposition method due to its use of perfect static load balance. For applications with a smaller ratio of macroparticles to grid points, the domain decomposition method exhibits better scalability and faster speed. Application of the particle-field decomposition scheme to high-resolution macroparticle-dominant parallel beam dynamics simulation for a future light source linear accelerator is presented as an example.     

PC机基于EPP接口的实验硬件扩展总线

辅助学生在微机原理与接口控制学习中建立从原理到应用的认识是微机接口实验应达到的首要目的。现有的实验系统,大多通过上位PC机与实验箱上的代理或目标处理器通信实现对实验电路的控制,其控制方式是间接的,不直观,且增加额外成本和环节,在实际工程应用中也很少使用。EPP直接接口的微机原理实验系统去除以往实验系统上的处理器,采用简单数字电路实现基于PC机EPP并口的标准总线扩展并连接实验电路,使控制访问更为直接,也更易与实际微机控制应用相结合。通过扩展总线接口还能连接学生自己设计的实验电路,有利于开放性实验的设计与实现。基于这一思想设计构造的实验系统经实际教学验证,效果良好。     

ETL多数据流并行抽取中监控的研究与设计

对ETL多数据流并行抽取进行内容与流程的监控,并对各监控模块作了详细分析与设计,监控提高了数据抽取的效率与数据质量,有效保证了系统稳定运行与数据抽取的顺利进行。     

机群环境下傅立叶变换的并行算法研究

并行FFT是解决大数据量FFT运算耗时过久的重要途径,在PC机群上实现并行FFT是一种低成本、高效率的解决方案。本文讨论了PC机群环境下MPI并行FFT实现,并利用建立的平台,对并行算法进行了测试,得出了一些有意义的结论和方法。     

Embedding multi-dimensional meshes into twisted cubes

The twisted cube is an important variant of the most popular hypercube network for parallel processing. In this paper, we consider the problem of embedding multi-dimensional meshes into twisted cubes in a systematic way. We present a recursive method for embedding a family of disjoint multi-dimensional meshes into a twisted cube with dilation 1 and expansion 1. We also prove that a single multi-dimensional mesh can be embedded into a twisted cube with dilation 2 and expansion 1. Our work extends some previously known results.     

Relating the power of the Multiple Associative Computing (MASC) model to that of reconfigurable bus-based models

The MASC (Multiple ASsociative Computing) model is a multi-SIMD model that uses control parallelism to coordinate the interaction of data parallel threads and supports associative SIMD computing on each of its threads. There have been a wide range of algorithms developed for this model. Research on using this model in real-time system applications and building a scalable MASC architecture is currently quite active. In this paper, we present simulations between MASC and reconfigurable bus-based models, e.g., various versions of the Reconfigurable Multiple Bus Machine (RMBM). Constant time simulations of the basic RMBM by MASC and vice versa are obtained. Simulations of the segmenting RMBM, fusing RMBM, and extended RMBM by MASC in non-constant time are also discussed. By taking advantage of previously established relationships between RMBM and two other popular parallel computational models, namely, the Reconfigurable Mesh (RM) and the Parallel Random Access Machine (PRAM), we extend our simulation results to further categorize the power of the MASC model in relation to RM and PRAM.     

A parallel direct solver for the self-adaptive hp Finite Element Method

In this paper we present a new parallel multi-frontal direct solver, dedicated for the hp Finite Element Method (hp-FEM). The self-adaptive hp-FEM generates in a fully automatic mode, a sequence of hp-meshes delivering exponential convergence of the error with respect to the number of degrees of freedom (d.o.f.) as well as the CPU time, by performing a sequence of hp refinements starting from an arbitrary initial mesh. The solver constructs an initial elimination tree for an arbitrary initial mesh, and expands the elimination tree each time the mesh is refined. This allows us to keep track of the order of elimination for the solver. The solver also minimizes the memory usage, by de-allocating partial LU factorizations computed during the elimination stage of the solver, and recomputes them for the backward substitution stage, by utilizing only about 10% of the computational time necessary for the original computations. The solver has been tested on 3D Direct Current (DC) borehole resistivity measurement simulations problems. We measure the execution time and memory usage of the solver over a large regular mesh with 1.5 million degrees of freedom as well as on the highly non-regular mesh, generated by the self-adaptive hp$hp$-FEM, with finite elements of various sizes and polynomial orders of approximation varying from p=1$p=1$ to p=9$p=9$. From the presented experiments it follows that the parallel solver scales well up to the maximum number of utilized processors. The limit for the solver scalability is the maximum sequential part of the algorithm: the computations of the partial LU factorizations over the longest path, coming from the root of the elimination tree down to the deepest leaf.     

Kinematics, pattern recognition, and motion control of mask-panel alignment system

Kinematics, pattern recognition, and motion control methods are investigated as an integrated approach for a mask-panel visual alignment system, which consists of the vision system to extract the mask-panel misalignment and the stage control system to compensate for it. First, a 4PPR alignment mechanism is considered and the inverse kinematic solution is found out to define the relationship between the mask-panel misalignment and the displacements of active joints. Then, a fast alignment mark recognition algorithm is proposed in terms of the geometric template matching (GTM), which is specific to simply shaped patterns but computationally much efficient compared with general correlation-based matchings. Finally, the kinematic solution and the developed vision algorithm are incorporated to implement a two-stage position-based visual servo, where both original fine images and reduced coarse ones are utilized together and the GTM plays a crucial role in achieving a near real-time visual feedback. Experimental results are shown to demonstrate the effectiveness of the GTM-based two-stage alignment control.