Papers by Keyword: High Performance Computing

Abstract: Thread level speculation has been proposed and researched to parallelize traditional sequential applications on homogeneous multi-core architecture. In this paper, a heterogeneous multi-core hardware simulation system is present, which provides with TLS execution mechanism. With a novel TLS programming model and a number of new speculative tuning techniques, benchmark Gzip is parallelized from-3% to 195% on a four-core processor, and the speedup of the test benchmarks are 30%, 43% and 156%, respectively with arbitrary, hotspot and insight speculation.

Abstract: In the current worldwide basketball game and mass sports are gradually paid more attention. And the current basketball technology is also constantly improvement, in order to strengthen the basketball defense technology, meanwhile in order to ascend the basketball skills of basketball player, this paper analyze and measure the optimization defensive position of basketball movement on the basis of the basketball defense technology action conversion training. Based on the actual investigation and test of several basketball team to carry out the method in combination with the model simulation and calculation and summarized statistics and contrasted research for basketball team game data result, so as to analysis the application in action conversion technology and methods of the basketball defense technology, in combination of qualitative and quantitative way, put forward the action conversion way of basketball defense technology, provided effectively directed basis and advice for basketball player to improve game skills and defensive skills.

Abstract: MPI is one of the important standards in high performance computing. MPI performance is generally focused on collective communications. And FCA (Fabric Collective Accelerator) is a new method accelerating collective communications. Through high performance computing environment testing, this paper mainly analyses the result of FCA with shared memory and without share memory accelerating IBM Platform MPI, FCA's principle and integration between IBM Platform MPI and FCA. At the same time, this paper may be a good reference for high performance computing using FCA.

Abstract: We review the application of advanced numerical techniques such as adaptive mesh refinement, implicit time-stepping, multigrid solvers and massively parallel implementations as a route to obtaining solutions to the 3-dimensional phase field problem for coupled heat and solute transport during non-isothermal alloy solidification. Using such techniques it is shown that such models are tractable for modest values of the Lewis number (ratio of thermal to solutal diffusivities). Solutions to the 3-dimensional problem are compared with existing solutions to the equivalent 2-dimensional problem.

Abstract: Modern GPUs (graphical processing units) are a common source of processing power inmany supercomputers. Their performance derives from the highly parallel architecture that is em-ployed and have the benefit of low cost, temperature and power consumption. Two finite differencemodels have been implemented on GPU, a semi-implicit and an explicit algorithm, to numericallymodel a stratified shear layer, that needs fine meshes to be modelled accurately. The GPU modelswere shown to improve performance by factors of around 50x and 20x for the semi-implicit and ex-plicit models respectively.

Abstract: In this paper, electromagnetic dynamic characteristics of suspension system of middle-low speed maglev train are analyzed with finite element analysis (FEA) method based on the high-performance computing platform (HPC). The couple structure between F-type track and suspension magnet is meshed by pretension element. The dynamic characteristics of suspension system are simulated in three-dimensional model with 4 degrees of freedom motions condition. Both the numerical simulations and the actual force tests of suspension system are carried out with the same input. The result shows that the calculation accuracy of finite element analysis is high.

Abstract: The advent of cloud is drastically changing the High Performance Computing (HPC) application scenarios. Current virtual machine-based IaaS architectures are not designed for HPC applications. This paper presents a new cloud oriented storage system by constructing a large scale memory grid in a distributed environment in order to support low latency data access of HPC applications. This Cloud Memory model is built through the implementation of a private virtual file system (PVFS) upon virtual operating system (OS) that allows HPC applications to access data in such a way that Cloud Memory can access local disks in the same fashion.

Abstract: Currently, the best known algorithm for factoring RSA modulus is the General Number Field Sieve. Through the software optimized implementation of GNFS with RSA-768, we extracted nine main calculation components from the lattice sieve. Detail descriptions and comprehensive analysis of the properties about calculation, memory and communication to the nine components were given in this paper, which makes it possible to use of a variety of computing platforms, such as CPU, FPGA, CELL, and GPU etc, to accelerate the realization of GNFS.

Abstract: The growth of serial and High Performance Computing (HPC) applications presents the challenge of porting of scientific and engineering applications. A number of key issues and trends in High Performance Computing will impact the delivery of breakthrough science and engineering in the future. ONAMA was developed to cope with increasing demands for HPC. ONAMA, which means a new beginning, is a desktop based Graphical User Interface which is developed using C and GTK. It aims to satisfy the research needs of academic institutions. ONAMA is a comprehensive package, comprising of applications covering many engineering branches. ONAMA provides tools that have a close affinity with practical simulation, thus making the learning process for students more applied. Most of the software tools and libraries are open source and supported on Linux, thereby promoting the use of open source software. It also provides tools to the researchers to solve their day-to-day as well as long term problems accurately in lesser time. The Execution Model of ONAMA serves to execute engineering and scientific applications either in sequential or in parallel on Linux computing clusters.

Abstract: High-performance computer was developed to address the tight schedule and the high reliability problems. For the 100 trillion times supercomputer Dawning 5000A computing nodes A950r-F server's reliability problems, on the bases of the pre-use investigation and stress analysis of the type, the thesis described the reliability accelerated test of A950r-F server implementation process, and obtained a series of accelerated test data. Accelerated test results from the failure analysis give the corresponding improvement measures by the reliability accelerated test of the A950r-F server. The experiment proved the feasibility and effectiveness of the guidelines.