动态二进制翻译中热路径优化的软件实现

在动态二进制翻译中，热路径的识别和生成是提高二进制翻译器效率的重要环节。为了提高热路径预测的命中率，必须在程序的运行中搜集较为详细的信息，这必然增加系统的开销。因此，在准确率和开销之间做出权衡十分必要。该文在研究现有热路径算法的基础上，提出了一种改进的基于路径的热路径识别和优化算法，并对结果进行了分析。     

动态二进制翻译的优化

提出动态二进制翻译的两种优化方案:基本块和热路径;分析了从代码中抽取值得优化部分的详细过程;同时也给出针对这两种方案的一些优化方法.最后简单介绍了当前一些动态二进制翻译系统所采用的优化技术.     

基于龙芯3A处理器的跨平台二进制翻译器QEMU的优化研究

动态二进制翻译与优化技术推动了计算机体系结构的发展。以龙芯3A处理器为实现平台,在系统级模拟模式下,针对原二进制翻译器QEMU寄存器分配方案提出跨平台优化方法,同时引入热路径对动态二进制翻译过程中冗余代码进行优化。实验结果表明,经过优化后,在龙芯3A平台上通过QEMU翻译器运行Windows XP操作系统,其常用X86应用程序的运行速度得到可观提升。以上研究对基于龙芯平台的系统级跨平台应用程序的实用化具有深远的现实意义。     

基于热例程的动态二进制翻译优化

依据对系统级程序行为特性的观察,提出了一种基于热例程的动态二进制翻译优化方法。该方法以频繁执行的例程作为优化单位,通过块内和块间优化算法消除动态二进制翻译引入的冗余。相比基于踪迹的优化方法,该方法具有优化单位发现开销更小、代码区域更大、无重复翻译等优点,更适用于系统虚拟机中操作系统代码的优化。在跨平台系统虚拟机监控器ARCH-BRIDGE上的测试表明,通过对内核代码实施该优化方法,SPEC CPUINT 2006程序的效率提升了3.5%~14.4%,相比基于踪迹的优化,性能最大提升了5.1%。     

基于隐马尔可夫模型的热路径预测算法研究*

基于热路径的动态优化技术是动态二进制翻译器中提高软件运行效率的一种有效方法。如何利用基本块中已有的有限历史运行信息来识别热路径并提高它的预测命中率,同时保持计算开销没有增加是研究的重点。已有的热路径识别算法中基于模型进行预测的方法非常少,算法实现比较复杂。基于隐马尔可夫模型提出一种改进的热路径预测算法。由于状态转移序列惟一,该算法实现简单,可以提高热路径的命中率,在一定程度上改善动态二进制翻译器的性能。最后通过实验对所提出算法的有效性进行验证。     

一种基于动态二进制翻译的硬件支持Profile技术

传统的基于动态二进制翻译器的profiling策略分为3种:基于基本块、基于跳转边、基于路径跟踪。使用纯软件的profiling系统一般地说会带来平均30%的性能开销。如果在动态优化中得到硬件的支持,系统的整体性能将得到显著的提高。其中,软硬件协同设计中的难点,就是软硬件之间的通信开销和软硬件划分。该文针对动态二进制翻译中的优化阶段,使用一种硬件支持的运行是profile收集新方法来取代纯软件的profiling方法,把软硬件之间的通信开销降到最低,并以此来提高动态二进制翻译的整体性能。此方法可以在运行时准确地,并且以很小的开销收集Profile信息,从而更好的优化系统。     

动态二进制翻译的多线程并行优化研究

为了充分利用多核CPU来实现动态二进制翻译的并行化,研究了用多线程将翻译阶段和执行阶段并行优化的方法,提供了并行化系统的程序流程。并根据翻译与执行的时序及相关性,设计实现了一种超前翻译算法,它能够有效预测程序的执行路径,为翻译过程提供导向作用。实验结果表明,该优化方法提高了翻译缓存中基本块的命中率,使执行阶段尽量不被中断,进而提升了执行效率。     

动态二进制翻译器CrossBit的性能分析与评估

动态二进制翻译是广泛应用于虚拟机系统的一种二进制代码的翻译技术。动态二进制翻译由于拥有代码缓存、本地执行、代码块链接、动态热路径生成等优化技术的支持,有着很高的性能。CrossBit是一个多元多目标的动态二进制翻译系统,通过对CrossBit二进制翻译器的性能进行的研究,分析动态二进制翻译器性能提升中所必须解决的若干问题,并通过定量的分析总结了一些二进制翻译系统的在不同的配置和负载下系统优化手段的执行时策略。     

基于障碍物运动预测的移动机器人路径规划

针对移动机器人局部动态避障路径规划问题开展优化研究。基于动态障碍物当前历史位置轨迹,提出动态障碍物运动趋势预测算法。在移动机器人的动态避障路径规划过程中,考虑障碍物当前的位置,评估动态障碍物的移动轨迹;提出改进的D*Lite路径规划算法,大幅提升机器人动态避障算法的效率与安全性。搭建仿真验证环境,给出典型的单动态障碍物、多动态障碍物场景,对比验证了避障路径规划算法的有效性。     

基于Dijkstra算法的动态交通诱导技术及仿真

研究车辆行驶过程中的路径动态诱导问题,针对目前交通导航系统不能实时动态规划行驶路线的不足,结合自主研发的车载终端装置,通过对Dijkstra算法的改进及优化,提出了一个可应用于交通诱导过程的动态实时最优路径算法;基于该路径优化算法,车载终端装置可以通过接受交通控制中心的实时道路信息,不断调整车辆的行驶路线,最终实现行驶路线的全程动态优化;仿真实例证明:在实时交通信息的引导下,动态交通诱导技术保证了行驶路线的全程优化.     

A wave propagation approach for reduced dynamic modeling of distillation columns: Optimization and control

Reduced models enable real-time optimization of large-scale processes. We propose a reduced model of distillation columns based on multicomponent nonlinear wave propagation (Kienle 2000). We use a nonlinear wave equation in dynamic mass and energy balances. We thus combine the ideas of compartment modeling and wave propagation. In contrast to existing reduced column models based on nonlinear wave propagation, our model deploys a hydraulic correlation. This enables the column holdup to change as load varies. The model parameters can be estimated solely based on steady-state data. The new transient wave propagation model can be used as a controller model for flexible process operation including load changes. To demonstrate this, we implement full-order and reduced dynamic models of an air separation process and multi-component distillation column in Modelica. We use the open-source framework DyOS for the dynamic optimizations and an Extended Kalman Filter for state estimation. We apply the reduced model in-silico in open-loop forward simulations as well as in several open- and closed-loop optimization and control case studies, and analyze the resulting computational speed-up compared to using full-order stage-by-stage column models. The first case study deals with tracking control of a single air separation distillation column, whereas the second one addresses economic model predictive control of an entire air separation process. The reduced model is able to adequately capture the transient column behavior. Compared to the full-order model, the reduced model achieves highly accurate profiles for the manipulated variables, while the optimizations with the reduced model are significantly faster, achieving more than 95% CPU time reduction in the closed-loop simulation and more than 96% in the open-loop optimizations. This enables the real-time capability of the reduced model in process optimization and control.     

Hopping on Even Ground and Up Stairs with a Single Articulated Leg

This paper presents a method for generating gaits for a one-legged articulated hopping robot. A static optimization procedure produces the initial joint velocities for the flight phase, using the principle of conservation of angular momentum and assuming (nearly) passive flight. Two novel objective functions for this static optimization enable one to choose different gaits by simply changing a few parameters. A dynamic optimization procedure yields a solution for the flight trajectory that minimizes control effort. The stance phase (when the foot is touching the ground) becomes a standard two point boundary value problem, also solved with a dynamic optimization procedure. During the stance phase, the physical joint limitations, ground reaction forces, and the trajectory of the zero-moment point all constrain the solution. After these single-phase optimizations, a complete-cycle optimization procedure, incorporating both flight and stance phases, further reduces the control effort and balances the motion phases. In simulation, the leg hops on even ground and up stairs, exhibiting energy-efficient and intuitively satisfying gaits.     

Code Optimization Considerations in List Processing Systems

Code optimization is characterized as a time versus space tradeoff. Space optimizations are further decomposed into static and dynamic categories. Using this characterization, the optimization requirements of a list processing language such as LISP are examined. Scrutiny of the structure of programs written in such a language reveals that traditional code optimization techniques have little benefit. Instead, a collection of low-level time and static space optimizations is seen to lead to a potential decrease in space and execution time. Dynamic space optimization is also examined in the context of reducing the frequency of occurrence of garbage collection. Alternatively, some language extensions are proposed which reduce the amount of storage that needs to be allocated, and hence may result in a decrease in the frequency of garbage collection.     

Strategies for Source-to-Source Constant Propagation

Data-flow optimizations are usually implemented on low-level intermediate representations. This is not appropriate for source-to-source optimizations, which reconstruct a source level program after transformation. In this paper we show how constant propagation, a well known data-flow optimization problem, can be implemented on abstract syntax trees in Stratego, a rewriting system extended with programmable rewriting strategies for the control over the application of rules and dynamic rewrite rules for the propagation of information.     

Cache oblivious algorithms for nonserial polyadic programming

The nonserial polyadic dynamic programming algorithm is one of the most fundamental algorithms for solving discrete optimization problems. Although the loops in the nonserial polyadic dynamic programming algorithm are similar to those in matrix multiplication, the available automatic optimization techniques have little effect on this imperfect loop because of nonuniform data dependencies. In this paper, we develop algorithmic optimizations to improve the cache performance of the nonserial polyadic dynamic programming algorithm. Our algorithmic transformation takes advantage of the cache oblivious method by relaxing some dependencies in the standard iterative version. Based on the ideal cache model of the cache oblivious algorithm, the approximate bound of cache misses is given by . We also found that the optimized algorithm with the cache oblivious approach is more sensitive to conventional optimization techniques such as tiling. Experimental results on several platforms show that the optimized algorithms improve the cache performance and achieves speedups of 2–10 times.

An ACO algorithm for a dynamic regional nurse-scheduling problem in Austria

To the best of our knowledge, this paper describes the first ant colony optimization (ACO) approach applied to nurse scheduling, analyzing a dynamic regional problem which is currently under discussion at the Vienna hospital compound. Each day, pool nurses have to be assigned for the following days to public hospitals while taking into account a variety of soft and hard constraints regarding working date and time, working patterns, nurses qualifications, nurses’ and hospitals’ preferences, as well as costs. Extensive computational experiments based on a four week simulation period were used to evaluate three different scenarios varying the number of nurses and hospitals for six different hospitals’ demand intensities. The results of our simulations and optimizations reveal that the proposed ACO algorithm achieves highly significant improvements compared to a greedy assignment algorithm.     

An improved dynamic membrane evolutionary algorithm for constrained engineering design problems

Constrained global optimization is a highly important and challenging task in the field of optimization, and is embedded in many science and engineering optimizations. In this paper, an improved dynamic membrane evolutionary algorithm based on particle swarm optimization and differential evolution (IDMEA-PSO/DE) is proposed to solve constrained engineering design problems. The method combines the dynamic membrane structure of P systems and the PSO/DE search strategy. The performance of IDMEA-PSO/DE is tested on several well-known engineering design problems. The results of the simulation experiments show that the proposed algorithm is effective and outperforms other state-of-the-art-algorithms.     

MapReduce大数据处理平台与算法研究进展

本文综述了近年来基于MapReduce编程模型的大数据处理平台与算法的研究进展。首先介绍了12个典型的基于MapReduce的大数据处理平台,分析对比它们的实现原理和适用场景,抽象它们的共性。随后介绍基于MapReduce的大数据分析算法,包括搜索算法、数据清洗/变换算法、聚集算法、连接算法、排序算法、偏好查询、最优化算法、图算法、数据挖掘算法。将这些算法按MapReduce实现方式分类,分析影响这算法性能的因素。最后,将大数据处理算法抽象为外存算法,并对外存算法的特征加以梳理,提出了普适的外存算法性能优化方法的研究思路和研究问题,以供研究人员参考。具体包括优化外存算法的磁盘I/O,优化外存算法的局部性,以及设计增量式迭代算法。现有大数据处理平台和算法研究多集中在基于资源分配和任务调度的平台动态性能优化、特定算法并行化、特定算法性能优化等领域,本文提出的外存算法性能优化属于静态优化方法,是现有研究的良好补充,为研究人员提供了广阔的研究空间。