会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 11. 发明申请
    • Self-Optimizable Code
    • 自我优化代码
    • US20080222637A1
    • 2008-09-11
    • US12119784
    • 2008-05-13
    • Marc Alan Dickenson
    • Marc Alan Dickenson
    • G06F9/46
    • G06F11/3409G06F9/5016G06F11/3466G06F2201/865G06F2201/88
    • Methods, systems, and media to increase efficiency of tasks by observing the performance of generally equivalent code paths during execution of the task are disclosed. Embodiments involve a computer system with software, or hard-coded logic that includes reflexive code paths. The reflexive code paths may be identified by a software or hardware designer during the design of the computer system. For that particular computer system, however, one of the code paths may offer better performance characteristics so a monitor collects performance data during execution of the reflexive code paths and a code path selector selects the reflexive code with favorable performance characteristics. One embodiment improves the performance of memory allocation by selectively implementing a tunable, linear, memory allocation module in place of a default memory allocation module.
    • 公开了通过在执行任务期间观察一般等效的代码路径的性能来提高任务效率的方法,系统和媒体。 实施例涉及具有软件的计算机系统或包括反身代码路径的硬编码逻辑。 在计算机系统的设计期间,反身代码路径可由软件或硬件设计者识别。 然而,对于该特定计算机系统,代码路径之一可以提供更好的性能特性,因此监视器在执行反射码路径期间收集性能数据,并且代码路径选择器选择具有良好性能特性的反射码。 一个实施例通过选择性地实现可调谐线性存储器分配模块来代替默认存储器分配模块来提高存储器分配的性能。
    • 12. 发明授权
    • Memory leakage management
    • 内存泄漏管理
    • US07398369B2
    • 2008-07-08
    • US10976269
    • 2004-10-28
    • Marc Alan Dickenson
    • Marc Alan Dickenson
    • G06F12/02
    • G06F11/362G06F12/023G06F2212/1032G06F2212/1044Y10S707/99953Y10S707/99957
    • Methods, systems, and media to enhance memory leakage management by identifying a suspect allocation pattern during execution of a task, which may be indicative of memory leakage and implementing measures to protect against memory leakage based upon the suspect allocation pattern, are disclosed. More specifically, embodiments may detect a suspect allocation pattern by monitoring memory allocations and deallocations for tasks. The pattern of memory allocations and deallocations may then be analyzed to determine whether a suspect allocation pattern exists. For instance, the memory allocations and deallocations may be compared to determine whether there is an increasing net residual allocation left by the task after it has reached a quiescent runtime state. In some embodiments, a suspect allocation pattern exists if the total memory allocation for a task continues to rise after a pre-determined time period.
    • 公开了通过在执行任务期间识别可疑分配模式来增强内存泄漏管理的方法,系统和媒体,其可以指示存储器泄漏并且基于可疑分配模式来实施防止内存泄漏的措施。 更具体地,实施例可以通过监视用于任务的存储器分配和释放来检测可疑分配模式。 然后可以分析存储器分配和释放的模式以确定是否存在可疑分配模式。 例如,可以比较存储器分配和释放分配,以确定任务在达到静态运行时状态之后是否存在增加的净剩余分配。 在一些实施例中,如果任务的总存储器分配在预定时间段之后继续上升,则存在可疑分配模式。