An HLA/RTI Architecture Based on Multi-thread Processing

In order to improve the real-time performance of the real-time HLA(high level architecture) in the application of massive data communication volume,multi-thread processing was adopted,thread pool structure was introduced into the system,different threads to handle corresponding message queues was utilized to respond different message requests.Furthermore,an allocation strategy of semi-complete deprivation of priority was adopted,which reduces thread switching cost and processing burden in the system,provided that the message requests with high priority can be responded in time,thus improves the system's overall performance.The design and experiment results indicate that the method proposed in this paper can improve the real-time performance of HLA in distributed system applications greatly.

Using multi-threads to hide deduplication I/O latency with low synchronization overhead

Data deduplication, as a compression method, has been widely used in most backup systems to improve bandwidth and space efficiency. As data exploded to be backed up, two main challenges in data deduplication are the CPU-intensive chunking and hashing works and the I/0 intensive disk-index access latency. However, CPU-intensive works have been vastly parallelized and speeded up by multi-core and many-core processors; the I/0 latency is likely becoming the bottleneck in data deduplication. To alleviate the challenge of I/0 latency in multi-core systems, multi-threaded deduplication （Multi-Dedup） architecture was proposed. The main idea of Multi-Dedup was using parallel deduplication threads to hide the I/0 latency. A prefix based concurrent index was designed to maintain the internal consistency of the deduplication index with low synchronization overhead. On the other hand, a collisionless cache array was also designed to preserve locality and similarity within the parallel threads. In various real-world datasets experiments, Multi-Dedup achieves 3-5 times performance improvements incorporating with locality-based ChunkStash and local-similarity based SiLo methods. In addition, Multi-Dedup has dramatically decreased the synchronization overhead and achieves 1.5-2 times performance improvements comparing to traditional lock-based synchronization methods.

Dynamically Translating Binary Code for Multi-Threaded Programs Using Shared Code Cache

mc211vm is a process-level ARM-to-x86 binary translator developed in our lab in the past several years. Currently, it is able to emulate singlethreaded programs. We extend mc211vm to emulate multi-threaded programs. Our main task is to reconstruct its architecture for multi-threaded programs. Register mapping, code cache management, and address mapping in mc2llvm have all been modified. In addition, to further speed up the emulation, we collect hot paths, aggressively optimize and generate code for them at run time. Additional threads are used to alleviate the overhead. Thus, when the same hot path is walked through again, the corresponding optimized native code will be executed instead. In our experiments, our system is 8.8X faster than QEMU （quick emulator） on average when emulating the specified benchmarks with 8 guest threads.

On Multi-Thread Crawler Optimization for Scalable Text Searching

Web crawlers are an important part of modern search engines.With the development of the times,data has exploded and humans have entered a“big data era”.For example,Wikipedia carries the knowledge from all over the world,records the realtime news that occurs every day,and provides users with a good database of data,but because of the large amount of data,it puts a lot of pressure on users to search.At present,single-threaded crawling data can no longer meet the requirements of text crawling.In order to improve the performance and program versatility of single-threaded crawlers,a high-speed multi-threaded web crawler is designed to crawl the network hyper-scale text database.Multi-threaded crawling uses multiple threads to process web pages in parallel,combining breadth-first and depth-first algorithms to control web crawling.The practice project is based on the Python language to achieve multi-threaded optimization network hyper-large-scale text database-Wikipedia book crawling method,the project is inspired by the article on the Wikipedia article in the Big Data Digest public number.

Research and Review on the Java Multi-thread Programming and Its Further Development Tendency

In this paper, we conduct research on the Java multi-thread programming and its further development tendency. Multithreading mechanisms can run several programs at the same time, make the program run effi ciency becomes higher that also can overcome the problem of basic traditional programming language design while its design is the key to the realization of the synchronous thread. Multithreading is a mechanism that allows concurrent execution of multiple instruction stream in the program, each instruction stream is called a thread, independent from each other between each other. Thread is also known as a lightweight process, it have independent execution and process control. Our research starts from the analysis of the corresponding mechanism to enhance the performance that is innovative and meaningful.

Multi-threaded Fast Distance Field Computation Using Spatial Partition

This article describes three algorithms for distance field generation on triangulated model: brute force algorithm, single-threaded algorithm based on spatial partition and multi-threaded algorithm based on spatial partition. Spatial partition algorithm use equidistant network divide the bounding box into equal-sized cubes, calculates the maximum and minimum distances between the sample point and each of the small cubes,taking the minimum value from the maximum distance as the minimum distance from the sample point to the model named d1, comparing d1 with the distance from sample point to every little cube's minimum distance d2, if d1 <d2, the sample point's distance to all triangles inside this cube are greater than d1, skip this cube, otherwise, calculated the distance from the point to all the triangles intersect with the cube, then alternative d1 with the minimum value, circulate all small cubes intersect with the model. Comparing the calculation results, it can be seen that the algorithm about the multi-threaded distance field relative to the other two algorithms in computational speed is greatly improved especially for complex models.

Multi-threaded code generation from Signal program to OpenMP

The use of multi-core processors will become a trend in safety critical systems. For safe execution of multi- threaded code, automatic code generation from formal spec- ification is a desirable method. Signal, a synchronous lan- guage dedicated for the functional description of safety crit- ical systems, provides soundness semantics for determinis- tic concurrency. Although sequential code generation of Sig- nal has been implemented in Polychrony compiler, deter- ministic multi-threaded code generation strategy is still far from mature. Moreover, existing code generation methods use certain multi-thread library, which limits the cross plat- form executions. OpenMP is an application program inter- face （API） standard for parallel programming, supported by several mainstream compilers from different platforms. This paper presents a methodology translating Signal program to OpenMP-based multi-threaded C code. First, the intermedi- ate representation of the core syntax of Signal using syn- chronous guarded actions is defined. Then, according to the compositional semantics of Signal equations, the Signal pro- gram is synthesized to dependency graph （DG）. After par- allel tasks are extracted from dependency graph, the Signal program can be finally translated into OpenMP-based C code which can be executed on multiple platforms.

An Object-Oriented Multi-Thread Dialog Model

This paper presents a multi thread dialog model using extended state transition network model combined with an object oriented event model to specify and manage user interface. The model provides multi thread dialogs and the concurrent executing of user interface and application procedures. The forms of class based concurrency possible in the model emphasize the human computer interaction and exploit the concurrency across objects and within an object. Dialog descriptions using extended state transition networks, the automatic extraction of concurrency, the facilities for performing synchronization and communicating, generalization and specification, and the smooth transition from dialog design to programming language implementation are also given in this paper.

一种基于无锁队列的运行时多线程并行验证方法

运行时验证是一种动态的软件验证技术,主要包括使用形式化规约描述待验证性质、自动生成对应监控器以及监控器的插桩.然而现有的面向C语言程序的运行时验证技术存在一些局限性,主要体现在多监控器的情况下,现有的运行时验证工具只能使用串行的方式处理,这大大降低了验证效率.因此,本文在分析了形式化规约的基础上,提出了一种基于无锁队列的运行时多线程并行验证方法.方法在现有工具MOVEC上实现并在测试集mibench上插桩运行,并与相关工具ACC、AC++和串行机制下的MOVEC进行了对比实验.实验结果表明,本文所实现的基于无锁队列的运行时多线程并行算法可以在有多个监控器的情况下有效地对C语言程序进行并行的运行时验证,且并行验证算法的性能比串行验证算法提升了约83%.

基于FFmpeg多线程编码的智能交通监控系统设计

针对智能交通管理设备本身缺乏安全监管,传统视频监控延迟高、画质低、稳定性差的问题,提出一种基于FFmpeg的多线程编码视频流传输方案。通过FFmpeg调用h264_nvenc编码器,实现宏块行级的GPU多线程加速,降低编码延迟。使用Visual Studio 2019和QT15.5开发基于FFmpeg的音视频处理软件,对多路视频流进行封装、推流,并搭建Nginx流媒体服务器进行分发。通过实验表明,该系统整体的传输延迟低于1 s,且拥有良好的率失真特性,监控画面清晰、稳定性高,实现了对交通管理设备实时稳定的安全监控。

基于国产ARM架构CPU的导航卫星精密定轨解算效率优化方法

以国产飞腾CPU为例,讨论在国产ARM架构CPU基础上的导航卫星精密定轨解算效率优化方法。基于导航卫星精密定轨解算流程中钟差约化和法方程求逆耗时较多,分别利用多线程和OpenBlas对上述2个过程进行优化。结果表明,优化后解算效率大幅提升。钟差约化方面,采用100个测站32颗导航卫星进行解算时,原始单历元平均耗时1.105 s,优化后为0.188 s;法方程求逆方面,原始求逆平均耗时2 264 s,优化后仅需78 s。

基于OpenMP的航迹融合并行优化方法

针对组网雷达中分布式数据融合计算量随数据量增大急剧增加的问题及航迹融合中最耗时的航迹关联环节,提出一种基于OpenMP的多线程优化方法。通过OpenMP并行计算雷达航迹间的关联距离,提升内部算法运算速度,并将数据接收、融合处理和结果输出过程进行多线程处理,进一步加快外部数据交互速度,提升融合处理整体时间性能。以大容量目标场景为测试用例,评估处理时间和优化加速比。仿真结果表明,所提并行优化方法能够有效提升运算速度。

一种多线程概念约简算法

针对现有概念约简算法计算过程较为繁琐且执行效率低的问题,提出了一种多线程概念约简算法MTCR。MTCR以提高计算概念约简的效率为首要目标,在多核环境下利用多线程技术并行计算概念约简。首先,MTCR算法使用两个线程分别计算单个对象的对象代表概念集和单个属性的属性代表概念集;然后,将形式背景中的对象(属性)依次放入p个队列,并为每个队列创建线程;最后,通过多线程方式并行计算任意对象的对象代表概念集和任意属性的属性代表概念集,以及两类代表概念集的交集,进而构建代表概念矩阵,据此计算出所有概念约简。在MTCR算法中,使用多线程的每个阶段的数据相对独立,使得多线程之间不需要频繁的同步操作,从而减少了线程之间的竞争和等待。这样可充分有效地利用计算资源,大大提升算法的性能。UCI数据集和随机数据集上的实验表明:MTCR算法可以准确得到概念约简结果,在使用单线程情况下执行速度与串行概念约简算法SCR相近;当线程数不超过8时,线程数每增加1倍,MTCR算法执行速度可提高30%以上。

基于DLL和多线程的随机序列测试软件设计

随机序列在基于密码学的网络安全算法中被大量应用,在网络安全应用中扮演着重要的角色,随机序列的测试方法也在持续发展,为了提高随机序列的随机性测试效率,开展了随机序列测试软件设计;通过对美国NIST公布的SP 800-22随机序列测试标准的梳理,在原配套测试集工具包的基础上,随机序列测试软件采用动态链接库和多线程技术,对软件设计架构和测试模式上进行了创新,提高了测试执行效率和人机界面友好性,并通过真随机数序列和伪随机数序列两种方式对其随机性检验效果进行了验证;经实际应用验证,该测试软件设计架构开放、灵活、可配置,测试模式支持多种随机性检验项目的并发执行,测试结果支持量化呈现,提升了测试评估人员的工作效率,在工程实践中具有一定的参考价值。

基于LabWindows/CVI的高效高精度专用测试设备设计

针对某型号炮射导弹仪器舱的传统测试设备无法满足高效率批量检测的问题,设计了基于LabWindows/CVI虚拟仪器平台的高效高精度专用测试设备。该设备通过标准化的通信协议、模块化、多线程、ACCESS数据库等关键性技术实现了系统自检、设备计量、多工位自动测试和数据管理等功能。经过实验验证,该专用测试设备对同一产品重复性测试时误差不超过0.5%,具有良好的稳定性和准确性。单发产品测试时各项技术指标测试误差较小,并行测试时比传统测试设备缩短了62.857%的测试时间,有效解决了多工位自动化并行测试问题,提高了测试效率和精度,降低了测试成本,促进了国防事业的发展。

基于改进多智能体PPO的多无人机协同探索方法

采用多无人机对未知环境进行探索,可以提高探索任务的鲁棒性和执行效率。不同于启发式方法,多智能体深度强化学习方法可以省去人为制定规则的过程,将无人机作为智能体,通过与环境互动,自主习得更加有效的“规则”。搭建了多无人机多线程仿真环境,为多无人机协同训练提供环境,提出一种适应多线程环境的结合长短时循环神经网络(记忆)的共享多智能体近端策略优化(LSTM-MAPPO)方法,并在合作型LSTM-MAPPO方法的基础上增加了全局边界信息以增大每幕探索面积。数值实验结果表明:与现有的多智能体深度确定性策略梯度(MADDPG)方法相比,所提方法在训练后期连续动作下也能稳定收敛;相较于现有的LSTM-MAPPO方法,其最终获得的奖励稳定高于5000;对3种不同的仿真地图,训练完的网络在测试时能实现70%以上的稳定探索面积。

高速旋转设备智能感知系统搭建及数据上传方法研究

针对高速旋转设备的特点搭建了物理智能感知系统,重点对系统所产生的传感器数据如何高效上传进行了深入研究。高速旋转设备物理感知系统具有采集信号多、采集方式多及数据量大等特点,提出3种上传算法:多线程多行算法可工作于普通单片机系统,上传速率达到0.06~0.07 MB/s,可满足系统工作于趋势值自动采集方式下的上传要求;单进程分页算法要求上传硬件系统为工业级单片机或嵌入式系统,上传速度可达0.7~0.8 MB/s,且稳定,可以满足系统工作于手动采集方式下数据实时上传的要求;多进程并行算法的硬件要求为多运算器设备,上传速率取决于运算器数量,10运算器主机上传速率稳定在约3.4 MB/s,可满足多设备系统任何数据采集方式下数据实时上传的要求。文中所做分析可为旋转机械行业数字化发展提供切实可行的方案。

轻量化智慧乡村移动用图系统的设计与实现

为推进慧智慧乡村信息化、智能化建设进程,解决移动用图系统商业软件价格昂贵、系统庞大、硬件要求高等问题,设计一个轻量化的智慧乡村移动用图系统。基于Android开发平台,结合开源GDAL和移动GIS技术,研究移动GIS、影像金字塔的多线程快速加载、多线程矢量分层按区动态加载、线与节点双属性等关键技术,建设轻量化的智慧乡村移动用图系统。以智慧园区数据为例,研发了轻量化、安装方便、运行快捷、操作简单、展示直观的移动用图系统,取得了较好的应用效果。

基于四阶龙格-库塔法的多弹道并行计算研究

弹道计算是一类典型的计算密集型问题,在飞行器设计及航迹规划中起着至关重要的作用,对该问题的求解通常具有计算量大、耗时长等问题。为实现弹道解算加速,本文利用多线程并行计算技术设计了基于四阶龙格-库塔法的多弹道加速计算架构,通过任务划分将多弹道解算映射到不同的线程并调度至处理器内核。通过试验分析了多弹道计算关键因素的影响,试验结果验证了利用多核多线程并行加速架构可以在保证高精度要求下有效地提高多弹道计算效率,并且可应用于提高微分方程类问题的解算效率。

多线程多核服务器功耗模型及性能监控

针对现有服务器的CPU功耗准确性不足和多线程处理功耗负载不平衡问题,本文提出了一种多线程和多核服务器CPU功耗模型。该模型采用前后端分离的整体架构思想,根据服务器功耗模型实时获取CPU的线程数、频率及利用率。通过日志文件进行反馈多线程和多核功耗模型,达到最佳服务器负载平衡。仿真测试的结果表明,改进的服务器功耗模型在CPU利用率方面比传统服务器提高了6%,内存以及I/O通信方面也有明显改善,为多线程和多核服务器提供了一种新的功耗模型思路。