基于TCG技术的二进制翻译条件转移指令优化研究

在二进制翻译中引入TCG中间表示技术可以实现多目标平台之间的程序移植,同时可以更加方便地引入新型平台,解决新平台对主流平台的兼容性问题。然而由于原有的中间表示在翻译过程中影响了代码的关联度,生成的后端代码中存在较多冗余指令,影响翻译程序的执行效率。分析了指令优化可行性,针对条件跳转指令进行优化,通过指令预处理对中间表示进行改进,实现中间表示到后端代码生成由一对多翻译模式到多对多翻译模式的转变,采用指令归约技术,针对条件跳转指令的2种模式CMP-JX型与TEST-JX型,分别设计相应的优化翻译算法,并在开源二进制平台QEMU上实现。基于NPB-3.3和SPEC CPU 2006测试集进行了测试,与以前的翻译模式进行对比,优化后的代码膨胀率平均减少了14.62%,翻译程序运行速度提升了17.23%,验证了该优化方法的有效性。     

一种可实时审计与阻断运维指令的变电站移动堡垒机的设计

本文针对传统变电站运维手段中,难以保障运维操作指令的可靠性和安全性,故提具有对运维指令实时审计与阻断能力的电网移动堡垒机运维系统设计研究。利用移动堡垒机可规范电网企业变电站运维工作体系,移动堡垒机的硬件模块包括感知集成芯片单元、数据传输与融合设备单元与设备工况监测单元;软件包括指令审计、指令识别、高危指令阻断等功能。通过硬件单元与软件功能的设计,在变电站环境下实现了移动堡垒机的应用。在仿真环境开展测试,结果显示设计系统针对违规指令处理的响应时间更短、处置成功率更高,充分表明设计系统运维性能更强、可靠性更高。     

面向国产高性能加速器的LLVM编译器设计及优化

国防科技大学自主研制的高性能加速器采用中央处理器(CPU)+通用数字信号处理器(GPDSP)的片上异构融合架构,使用超长指令集(VLIW)+单指令多数据流(SIMD)的向量化结构的GPDSP是峰值性能主要支撑的加速核。主流编译器在密集的数据计算指令排布、为指令静态分配硬件执行单元、GPDSP特有的向量指令等方面不能很好地支持高性能加速器。基于低级虚拟器(LLVM)编译框架,在前寄存器分配调度阶段,结合峰值寄存器压力感知方法(PERP)、蚁群优化(ACO)算法与GPDSP结构特点,优化代价模型,设计支持寄存器压力感知的指令调度模块;在后寄存器分配阶段提出支持静态功能单元分配的指令调度策略,通过冲突检测机制保证功能单元分配的正确性,为指令并行执行提供软件基础;在后端封装一系列丰富且规整的向量指令接口,实现对GPDSP向量指令的支持。实验结果表明,所提出的LLVM编译架构优化方法从功能和性能上实现了对GPDSP的良好支撑,GCC testsuite测试整体性能平均加速比为4.539,SPEC CPU 2017浮点测试整体性能平均加速比为4.49,SPEC CPU 2017整型测试整体性能平均...     

基于EPIC技术的密码处理器体系结构研究与设计

以分组密码和摘要算法为研究对象,结合处理器体系结构的特点,研究了能够高效灵活实现多种分组密码和摘要算法的处理器体系结构。通过分析现有分组密码算法的运算和结构特点,从实现方式的灵活性和高性能角度出发,提出了一种基于显式并行指令计算结构且性能和灵活性达到了折衷的可编程密码微处理器体系结构。给出了系统的整体架构、可重构运算单元的设计方案及其指令系统的设计,以及用硬件描述语言Verilog实现后的硬件测试参数。     

基于VelociTI体系结构的DSP指令分配的实现

在设计基于VelociTI体系结构的数字信号处理器过程中,为了高速实现并行指令的分配,提出了一种基于该体系结构的指令分配方法:排序法。该方法结合决策树原理实现取指包指令并行性测试,并将处理器的功能单元按照一个规定的顺序排列,使每一个功能单元与执行包的某一个字段对应,将执行包中的指令根据译码的结果和功能单元的顺序进行重新排序,从而完成指令到功能单元的分配。仿真结果证明该方法是十分有效的。     

二进制翻译后端指令分级索引策略

随着硬件平台的多元化,软件兼容性问题日益突出,二进制翻译技术是解决软件兼容性问题的重要手段.鉴于二进制翻译系统大部分执行时间消耗在后端指令的查找和执行过程中,高效的指令索引策略可以减少系统的指令查找开销,提高系统的整体效率.在对二进制翻译系统后端指令局部性特征进行统计分析的基础上,设计了一种能充分挖掘现代计算机系统硬件性能的二进制翻译指令分级索引策略.该策略结合二进制翻译系统后端指令特殊的局部性特征,使用针对性的替换算法对后端指令进行缓存,降低了系统的指令查找开销.在引入了LIIS索引策略后,开源二进制翻译系统QEMU的后端指令查找时间减少了70％,整个系统执行效率提高了15％.     

动态二进制翻译中全寄存器直接映射方法

二进制翻译是不同体系结构之间软件移植的重要手段。体系结构和硬件环境上的差别,可以通过二进制翻译系统来弥补,在翻译过程中往往使用多条本地指令模拟一条目标指令,翻译代码规模随之显著增加,从而导致被翻译程序的执行效率下降。寄存器作为处理器和内存交换信息的重要存储部件,寄存器的模拟器方式对于程序的性能有着至关重要的影响。为了提高特定平台翻译后代码的执行效率,提出了在动态二进制翻译机制中使用全部寄存器直接映射方法,详细分析了二进制翻译中的上下文切换原理和寄存器访问范围,为异构平台之间寄存器直接映射提供方法指导。利用QEMU模拟器,把x86架构的8个通用寄存器全部的直接映射到MIPS架构的对应寄存器,在此基础上,进行大量的指令翻译规则的简化。实验数据表明,该方法可以有效简化指令翻译,降低代码膨胀率,使得SPEC CINT 2000测试程序在龙芯CPU上翻译后代码运行时间下降了30%-40%。     

在计算机中实现三角函数的硬件解算方法

针对目标跟踪中大量存在的三角函数运算,提出在火控计算机中建立三角函数解算单元,采用硬件实现三角函数解算的设想。而三角函数解算单元是一种可以完成正弦函数解算功能的函数芯片。它的设计核心是以角度值为地址,通过硬件查表计算,可对0到90度的正弦函数值进行查表。查表精度可设计到0.001°,最大查表时间即为函数值的解算时间。余弦查表计算是采用正弦同一芯片,可以根据相位差90度的关系进行查表。实践表明三角函数的硬件解算速度快,精度高,可以替代通常方法的函数解算。因此,如果在火控系统中,设计一个地址译码器电路,就可建立多个三角函数芯片解算单元,通过计算机数据指令,实现对多个三角函数值的并行计算,从而提高系统对多个目标的快速跟踪处理能力。     

有限域模乘专用指令设计

针对椭圆曲线密码算法中有限域模乘运算的需求,提出其专用模乘指令。利用指令域中的组参数实现算法多组模乘运算,通过对参数进行配置,使指令支持运算长度拓展,在模乘运算单元中实现Montgomery模乘算法,并设计素域和二进制域统一的硬件流水线,以及双域乘法器单元结构。实验结果表明,该有限域模乘指令和硬件运算单元具有较高的执行效率和较好的灵活性。     

降低协同设计虚拟机启动开销的译码后指令缓存技术

协同设计虚拟机采用动态二进制翻译实现不同体系结构间的二进制兼容,对源指令的翻译和处理影响了协同设计虚拟机的启动性能.研究发现,在一个采用解释执行和翻译相结合的协同设计虚拟机中,处理非热点代码的解释执行是虚拟机启动开销的主要来源.发现了协同设计虚拟机中的解释例程局部性,并提出了一种硬件译码后指令缓存结构DICache(d...     

Sonification of guidance data during road crossing for people with visual impairments or blindness

In the last years several solutions were proposed to support people with visual impairments or blindness during road crossing. These solutions focus on computer vision techniques for recognizing pedestrian crosswalks and computing their relative position from the user. Instead, this contribution addresses a different problem; the design of an auditory interface that can effectively guide the user during road crossing. Two original auditory guiding modes based on data sonification are presented and compared with a guiding mode based on speech messages.Experimental evaluation shows that there is no guiding mode that is best suited for all test subjects. The average time to align and cross is not significantly different among the three guiding modes, and test subjects distribute their preferences for the best guiding mode almost uniformly among the three solutions. From the experiments it also emerges that higher effort is necessary for decoding the sonified instructions if compared to the speech instructions, and that test subjects require frequent ‘hints’ (in the form of speech messages). Despite this, more than 2/3 of test subjects prefer one of the two guiding modes based on sonification. There are two main reasons for this: firstly, with speech messages it is harder to hear the sound of the environment, and secondly sonified messages convey information about the “quantity” of the expected movement.     

Fast and flexible tracepoints in x86

Tracing is often the most effective technique for analyzing the performance of complex multithreaded applications. This paper presents an improvement on existing techniques for dynamic tracepoint insertion. To add a tracepoint, the technique inserts a jump at the tracing point, possibly replacing several shorter instructions. This jump embeds trap instructions inside its offset at the address of every replaced instruction. This makes the jump thread safe if any thread is about to execute a replaced instruction. It also makes it jump safe if a jump landing pad is at one of the replaced instructions. In both cases, a trap will be raised, and the thread can be redirected to the out-of-line equivalent instruction. The use of a jump instead of a trap to execute the tracepoint improves the performance of the execution. It also adds the flexibility to place the tracepoint at almost any instruction, since multiple instructions can be replaced atomically and safely. The downside of this technique is the increased memory usage, since it requires unaligned allocations with high external fragmentation.     

基于动态数据流分析的虚拟机保护破解技术

由于虚拟机采用虚拟化技术和代码混淆技术,采用传统的逆向分析方法还原被虚拟机保护的算法时存在较大困难。为此,提出一种基于动态数据流分析的虚拟机保护破解方法。以动态二进制插桩平台Pin作为支撑,跟踪记录被虚拟机保护的算法在动态执行过程中的数据流信息,对记录的数据流信息进行整理分析,获取虚拟机指令的解释执行轨迹,还原程序的控制流图,根据轨迹信息对数据生成过程进行分层次、分阶段还原,并由分析人员结合控制流图和数据生成过程进行算法重构。实验结果证明,该方法能够正确还原程序的控制流和数据生成过程,辅助分析人员完成被保护算法的重构。     

Mutual learning of multi-consciousness agents,including an ego for an autonomous vehicle

In this paper, we propose a multi-agent learning system for the control of an intelligent robot, based on a model of the human consciousnesses, including the ego. We pay attention to the intelligent learning processes of human beings. We try to give a robot a high learning ability by modeling the roles of the human consciousnesses, including the ego. In most ordinary methods, the instructions for learning are given from outside the system only. In the proposed method, the instructions are given not only from outside, but also from inside (from other agents in the system). Therefore, the robot can learn efficiently because it has more instructions than usual. The learning is also more flexible, since an agent learns by instructions from other agents while the learning agent and one of the instructing agents exchange roles according to changes in the environment. We experimentally verified that the proposed method is efficient by using an actual robot.     

DiffServ for differentiated reliability in meshed IP/WDM networks

The protection of service delivery in an integrated network is fast becoming a critical issue. In previous work, we proposed two protection methods, DiffServ and DiffProtect and examined, on a small network, whether DiffServ can be used as a protection mechanism against failures in IP/WDM networks. The results showed that, on average, DiffServ can provide better QoS guarantees than its optical counterpart, DiffProtect. We also found that DiffProtect may perform better than DiffServ under some extreme failure situations so that DiffServ cannot totally replace optical protection in a network but might be able to do so to some degree. We extend these results to larger networks and more complex combinations of protection techniques. We study in this paper a 4-node linear topology and a small 6-node mesh network which can be protected by DiffServ, DiffProtect or even a combination of both (MixProtect). Results show that replacing optical protection by DiffServ cannot be done randomly and that the number and position of links protected by DiffServ versus DiffProtect must be carefully chosen in order to guarantee adequate performance to all traffic types. We explain how this can be achieved with current technology. We also show that the cost of protection using DiffServ is significantly less than that of current methods.     

PrivaSIP: Ad-hoc identity privacy in SIP

In modern and future networks that belong to different providers, multimedia protocols will have to operate through multiple domains. In such an environment security is considered a crucial parameter; this is true especially for privacy since not all domains can be considered trusted beforehand in terms of personal data protection. Probably the most promising protocol for multimedia session management is SIP. While SIP is popular and a lot of research has been conducted, it still has some security issues, one of which is related to privacy and more particularly the protection of user identities (IDs). In the general case everybody can reveal the communicating parties IDs by simply eavesdropping on the exchanged SIP messages. In this paper we analyze the lack of user ID protection in SIP and propose two solutions; in the first the ID of the caller is protected while in the second both IDs of the caller and the callee are protected. Our work also includes performance results and extensive comparison with similar methods. The most significant advantage of our method is that it can assure user ID protection even when SIP messages are transmitted through untrusted SIP domains before reaching the Home Domain of the user or another trusted domain. Moreover, it does not require from the SIP Proxy server to maintain state information for exchanged SIP requests and respective responses.     

Suggestions on the applicability of visual instructions with see-through head mounted displays depending on the task

Task instructions have traditionally been communicated orally in many fields. However, recently more and more wearable displays, such as the see-through head mounted displays (HMDs) have been developed, and some studies have provided ideas on applying visual instruction using these new interfaces to particular situations. However, in some cases, where instructions are communicated amongst the workers, the data is not sufficient for field workers to choose the best way of communicating instructions depending on the situation. Thus, this study aims to clarify the cases in which it is effective to apply visual instructions with HMDs, and to provide information that suggests the applicability of such visual instructions instead of or in addition to the traditional auditory instructions in different situations. These suggestions will be a useful reference for workers in safety-critical fields, helping them make better decisions about whether, when, and where to introduce the new method of instructions. It will also address some of the unsolved problems in the field, such as errors, low efficiency, and discomfort in communication.     

Suggestions on the applicability of visual instructions with see-through head mounted displays depending on the task

Task instructions have traditionally been communicated orally in many fields. However, recently more and more wearable displays, such as the see-through head mounted displays (HMDs) have been developed, and some studies have provided ideas on applying visual instruction using these new interfaces to particular situations. However, in some cases, where instructions are communicated amongst the workers, the data is not sufficient for field workers to choose the best way of communicating instructions depending on the situation. Thus, this study aims to clarify the cases in which it is effective to apply visual instructions with HMDs, and to provide information that suggests the applicability of such visual instructions instead of or in addition to the traditional auditory instructions in different situations. These suggestions will be a useful reference for workers in safety-critical fields, helping them make better decisions about whether, when, and where to introduce the new method of instructions. It will also address some of the unsolved problems in the field, such as errors, low efficiency, and discomfort in communication.     

基于云计算背景下的计算机网络安全研究

云计算可以用于商业计算方法,它可以使计算指令发给很多计算机组建的资源池上,使用客户可以按照自己的去求得到想要的计算方式、存储空间和信息服务,客户可以从逻辑底层硬件、网络协议和软件中更好的解脱出来。我们在使用计算机技术中,肯定存在信息安全问题,所以我们需要加强计算机网络信息安全及防护。同时也要深入探讨和分析计算机网络防护以及安全。因此,本文主要就云计算背景下环计算机网络安全进行研究和讨论。也会主要说明做好计算机网络安全防范的重要性,希望可以在保证资源信息共享的情况下,使运行稳定,还可以使用户的生活和工作更加便捷。