CUDA平台下LISP2算法垃圾收集并行研究

为了提高垃圾收集效率,降低垃圾收集耗费时间,提出一种基于LISP2算法的并行节点复制垃圾收集算法,给出了在CUDA环境下该算法的实现。实验结果显示,该算法在CUDA环境下能有效提高垃圾收集效率。     

片上Trace辅助的嵌入式多核程序分析与调优

YHFT-QDSP是一款多核处理器,TraceDo是其实时片上追踪调试系统。本文首先将串行二维快速傅立叶变换算法(2D-FFT)并行化,映射于该多核处理器;并基于可视化的TraceDo系统,从多核程序同步及核间数据传输两个方面深入分析了并行2D-FFT程序行为;而后提出并实现了两种性能优化(调优)方案。实例研究表明,TraceDo系统可有效地提高嵌入式多核程序的开发效率。     

基于高清编码的自适应Slice划分算法

对于多核高清视频实时编码系统,提出一种自适应Slice划分算法。该算法基于码率控制和熵编码复杂度模型,通过Intra预测得到当前编码图像的纹理复杂度分布,预测编码图像的计算复杂度分布,通过自适应Slice划分实现多核间计算复杂度均匀分配,从而提高多核并行编码效率。实验结果表明,与固定宏块数的Slice划分算法相比,该算法能更有效地提高并行加速比。     

FLSP:一个高效的系统级垃圾收集算法

垃圾收集是Java操作系统的核心功能,它直接影响到整个系统效率。现代Java操作系统中使用的垃圾收集算法普遍还是沿用应用程序级的垃圾收集算法。应用程序级垃圾收集算法的优化主要面向于普通的Java虚拟机。而Ja-va操作系统与Java虚拟机相比有更高的操作权限和更灵活的资源管理策略,如何利用这些特点和权限来提高垃圾收集算法的效率是以前的垃圾收集算法所没有考虑的。本文分析了操作系统下内存管理和垃圾收集的特点,在JUnicorn操作系统上,利用操作系统平台提供的便利,设计并实现了一个高效的系统级垃圾收集算法FLSP。测试数据表明,在操作系统级别,这种垃圾收集算法能够提高13%的系统性能,并且使垃圾收集的停顿时间缩短50%。     

基于OpenMP多核架构下并行蚁群算法研究

研究了一种基于OpenMP技术的多核架构下并行蚁群算法,通过在TSP问题中的实验表明,该算法易于操作,而且充分利用了多核处理器并行计算的优势,提高了算法的运行效率。     

一种改进的大素数乘法的设计与实现

RSA算法作为应用较为广泛的非对称加密算法,经过蒙哥马利模乘等算法的优化后主要基于有限域运算中大数的加法运算和乘法运算,数位规模通常在1024位甚至更高。大数的乘法运算随着参与运算位数的增加会导致RSA算法的运行时间效率下降。随着多核处理器架构的普及,如何在多核多线程并行运算背景下提高RSA算法效率就成为解决RSA算法性能瓶颈的关键。本文通过多核并行运算背景下分析大数乘法算法从而提出一种改进的适应多核运算的大数相乘算法,依靠此算法提高RSA算法和大规模科学计算中高精度浮点数运算效率。     

多核处理平台上任务图模型的并行调度策略研究

凭借着高性能,低功耗的特性,多核处理器已经占据了目前的主要市场.提出一种多核处理平台上基于任务图模型的调度策略.建立了多核平台上任务图的空间与时间并行调度模型;针对任务图的空间并行与时间并行调度模型提出了并行节点合并、分配的优化算法与流水线并行的优化算法.最后,提出将优化的空间与时间并行调度技术相结合的并行调度策略.通过实验验证,本文提出的算法比其他多核并行调度算法降低了处理器核心间的通信与同步开销,提高了系统的计算效率与吞吐量.     

基于通用多核处理器平台的并行基因表达式编程算法

基因表达式编程(Gene Expression Programming, GEP)是一种计算量大且通用性强的新型进化算法,其传统计算形式不能充分利用目前主流的多核处理器。为提高算法效率,提出了基于通用多核处理器平台的并行基因表达式编程算法(Parallel Gene Expression Programming Based on General Multi-core Processor, PGEP-MP)。主要工作包括:O)分析通用多核处理器平台下并行基因表达式编程算法的机理;(2)利用MPI和()pcnMP混合编程模型设计基于通用多核处理器平台的基因表达式编程算法的粗粒度与细粒度相结合的并行模型;(3)提出改进PEEP-MP算法效率的进化策略;(4)通过对函数挖掘和分类的实验证明,PEEP-Ml〕算法提高了函数挖掘和分类的效率,在并行双核处理器数为4的情况下,PEEP-MP的平均并行加速比分别是传统GEP算法的4. 22倍和 4. 06倍。     

基于功能和数据划分的H.264并行解码算法

针对H.264视频解码算法的并行模块选择、划分及解码速度优化等问题,面向TilePro64多核平台,提出一种可扩展的H.264并行解码算法。对该算法的内部功能模块进行整合和划分,根据核间数据的依赖关系,动态分配功能模块及优化算法并行效率。实验结果表明,该算法在解码效率、多核并行程度、解码时延等方面均有较好性能,相比传统并行解码算法,其并行加速比提高约25%。     

多核处理器上的并行联机分析处理算法研究

近年来,计算机硬件技术获得了很大发展,尤其是大内存和多核,但算法效率并没有随着硬件技术的发展而提高,根本原因是没有充分利用CPU缓存以及单线程程序设计的局限性。在联机分析处理领域,数据方体计算是一个重要而又耗时的操作,因此如何提高数据方体的计算效率是该领域的一个研究难点。探讨了基于多核CPU特征的并行立方体算法,提出了MT-Multi-Way(multi-threading multi-way)和MT-BUC(multi-threading bottom-up computation)算法。该算法通过有效的数据划分和多线程协作,避免了Cache竞争,并确保了负载均衡,获得了近似线性加速比。以上述算法为基础,提出了处理立方体算法的多核框架,包括数据划分策略及递归算法的多核处理,指导立方体算法的并行化。     

A novel processor architecture with exact tag-free pointers

Along with the success of Java, garbage collection advances to embedded and real-time systems. In this paper a novel RISC processor architecture that lays the foundation for efficient real-time garbage collection in hardware and guarantees pointer integrity at the machine-code level is proposed. This article summarizes the most commonly used garbage collection methods and points out problems of contemporary software-based solutions to motivate architectural support for garbage collection.     

FeGC: An efficient garbage collection scheme for flash memory based storage systems

NAND flash memory is a promising storage media that provides low-power consumption, high density, high performance, and shock resistance. Due to these versatile features, NAND flash memory is anticipated to be used as storage in enterprise-scale systems as well as small embedded devices. However, unlike traditional hard disks, flash memory should perform garbage collection that consists of a series of erase operations. The erase operation is time-consuming and it usually degrades the performance of storage systems seriously. Moreover, the number of erase operations allowed to each flash memory block is limited. This paper presents a new garbage collection scheme for flash memory based storage systems that focuses on reducing garbage collection overhead, and improving the endurance of flash memory. The scheme also reduces the energy consumption of storage systems significantly. Trace-driven simulations show that the proposed scheme performs better than various existing garbage collection schemes in terms of the garbage collection time, the number of erase operations, the energy consumption, and the endurance of flash memory.     

A highly effective partition selection policy for object databasegarbage collection

We investigate methods to improve the performance of algorithms for automatic storage reclamation of object databases. These algorithms are based on a technique called partitioned garbage collection, in which a subset of the entire database is collected independently of the rest. We evaluate how different application, database system, and garbage collection implementation parameters affect the performance of garbage collection in object database systems. We focus specifically on investigating the policy that is used to select which partition in the database should be collected. Three of the policies that we investigate are based on the intuition that the values of overwritten pointers provide good hints about where to find garbage. A fourth policy investigated chooses the partition with the greatest presence in the I/O buffer. Using simulations based on a synthetic database, we show that one of our policies requires less I/O to collect more garbage than any existing implementable policy. Furthermore, that policy performs close to a locally optimal policy over a wide range of simulation parameters, including database size, collection rate, and database connectivity. We also show what impact these simulation parameters have on application performance and investigate the expected costs and benefits of garbage collection in such systems     

Scheduling of hard real-time garbage collection

Automatic memory management or garbage collection greatly simplifies development of large systems. However, garbage collection is usually not used in real-time systems due to the unpredictable temporal behavior of current implementations of a garbage collector. In this paper we propose a real-time garbage collector that can be scheduled like a normal real-time thread with a deadline monotonic assigned priority. We provide an upper bound for the collector period so that the application threads will never run out of memory. Furthermore, we show that the restricted execution model of the Safety Critical Java standard simplifies root scanning and reduces copying of static data. Our proposal has been implemented and evaluated in the context of the Java processor JOP.     

Schedulable garbage collection in CLI virtual execution system

Virtual software execution environment, known as Virtual Machine (VM), has been gaining popularity through Java Virtual Machine (JVM) and Common Language Infrastructure (CLI). Given their advantages in portability, productivity, and safety, etc., applying VM to real-time embedded systems can leverage production cost, fast time-to-market, and software integrity. However, this approach can only become practical once the VM operations and application tasks are made schedulable jointly. In this paper, we present a schedulable garbage collection algorithm applicable on real-time applications in CLI virtual machine environment. To facilitate the scheduling of real-time applications and garbage collection operations, we make the pause time due to garbage collection controllable, and the invocation of garbage collection predictable. To demonstrate the approach, a prototype for a schedulable garbage collection has been implemented in CLI execution environment. The garbage collection is carried out by a concurrent thread while meeting a targeted pause time and satisfying the memory requests of applications. A cost model of garbage collection is established based on measured WCET such that the execution time and overhead of garbage collection operations can be predicted. Finally, we illustrate a joint scheduling algorithm to meet the time and memory constraints of real-time systems.     

Global garbage collection for distributed heap storage systems

We present a garbage-collection algorithm, suitable for loosely-coupled multiprocessor systems, in which the processing elements (PEs) share only the communication medium. The algorithm is global, i.e., it involves all the PEs in the system. It allows space compaction, and it uses a system-wide marking phase to mark all accessible objects where a combination of parallel breadth-first/depth-first strategies is used for tracing the object-graphs according to a decentralized credit mechanism that regulates the number of garbage collection messages in the system. The credit mechanism is crucial for determining the space requirement of the garbage-collection messages. Also a variation of this algorithm is presented for systems with high locality of reference. It allows each PE to perform first its local garbage collection and only invokes the global garbage collection when the freed space by the local collector is insufficient.     

分代算法在KVM中的实现

在嵌入式系统中，需要分析什么样的垃圾回收机制能够适应嵌入式系统的特点，满足系统对于稳定性和性能的要求，分析了嵌入式Java虚拟机KVM垃圾回收机制的特点，所采用的垃圾收集算法，以及垃圾收集器在KVM中是如何工作的，找出存在的缺陷和不足，并在mark-sweep-compact算法的基础上实现了分代算法。     

Reference count garbage collection

A garbage collection algorithm that permits a reference count storage reclamation scheme to collect circularly linked inaccessible structures is presented. The algorithm requires no additional information beyond that required by a reference count scheme. In particular, it does not require the garbage collector to be able to find pointers outside the heap. The algorithm is most useful for augmenting reference count storage reclamation systems and for implementing storage management systems on top of languages that do not provide their own. It is, however, considerably less efficient in space and time than conventional garbage collection systems.     

Real-Time Garbage Collection for a Multithreaded Java Microcontroller

We envision the upcoming of microcontrollers and systems-on-chip that are based on multithreaded processor kernels due to the fast context switching ability of hardware multithreading. Moreover we envision an extensive market for Java-based applications in embedded real-time systems. This paper introduces two new garbage collection algorithms that are dedicated to real-time garbage collection on a multithreaded Java microcontroller. Our garbage collector runs in a thread slot in parallel to real-time applications. We show that our algorithms require only about 5–10% of the processor time for an effective garbage collection concerning our real-time benchmarks.     

The active deallocation of objects in object-oriented systems

In object-oriented systems, it is oftern useful for objects to be allowed to carry out some action before they are deallocated. This can be done by defining a destroy method in the object's class, and arranging for the memory system to send a message invoking this method immediately before deallocating the object. This allows resources associated with the object to be returned to the system, limited cross-language garbage collection, and other, more complex, behaviour. During the execution of the destroy method it is possible for new references to objects to be created. Care must be taken that the garbage collection does not erroneously free such objects. Algorithms are presented to implement destroy methods in systems using reference counting and mark-scan garbage collection techniques. Properties that are desirable in such systems are also discussed.