Methods and apparatus to implement annotation based thunking
    1.
    发明申请
    Methods and apparatus to implement annotation based thunking 失效
    实施基于注解的方法和装置

    公开(公告)号:US20070234286A1

    公开(公告)日:2007-10-04

    申请号:US11440850

    申请日:2006-05-25

    IPC分类号: G06F9/44 G06F9/45

    CPC分类号: G06F9/455

    摘要: Methods and apparatus to implement annotation based thunking are disclosed. An example method comprises locating a parameter of a function, the parameter to be passed as a pointer if a size of the parameter is greater than a threshold and to be passed as data if the size of the parameter is not greater than the threshold, and adding an annotation record for the parameter to a byte code image file containing byte code for the function.

    摘要翻译: 披露了实现基于注解的拆分的方法和装置。 一种示例性方法包括:如果参数的大小大于阈值,则定位要作为指针传递的参数的参数,并且如果参数的大小不大于阈值则作为数据传递;以及 将该参数的注释记录添加到包含该函数的字节码的字节码图像文件中。

    Compiling and translating method and apparatus
    2.
    发明申请
    Compiling and translating method and apparatus 审中-公开
    编译方法和装置

    公开(公告)号:US20090106744A1

    公开(公告)日:2009-04-23

    申请号:US10576907

    申请日:2005-08-05

    IPC分类号: G06F9/45

    CPC分类号: G06F8/447

    摘要: Methods and apparatus are described to compile and translate source code. In some embodiments, source code is compiled into source binary code for a source platform; an annotation section associated with the source binary code is generated, wherein the annotation section comprises an annotation for a scope, the scope comprising at least one block of the source binary code having at least one attribute to aid a translator optimization. If the scope comprises a plurality of blocks, the blocks have consecutive addresses with each other and have the at least one attribute in common. In the embodiments, the source binary code is further translated into target binary code for a target platform by utilizing the annotation section.

    摘要翻译: 描述了编译和翻译源代码的方法和设备。 在一些实施例中,源代码被编译为源平台的源二进制代码; 生成与源二进制代码相关联的注释部分,其中注释部分包括范围的注释,所述范围包括具有至少一个属性的源二进制代码的至少一个块以辅助翻译器优化。 如果范围包括多个块,则这些块具有彼此的连续地址并且具有共同的至少一个属性。 在实施例中,源二进制码通过利用注释部分进一步转换为目标平台的目标二进制代码。

    Redundant store-load instruction elimination based on stack location insensitive sequences
    3.
    发明申请
    Redundant store-load instruction elimination based on stack location insensitive sequences 审中-公开
    基于堆栈位置不敏感序列的冗余存储加载指令消除

    公开(公告)号:US20060218542A1

    公开(公告)日:2006-09-28

    申请号:US11091239

    申请日:2005-03-28

    IPC分类号: G06F9/45

    CPC分类号: G06F8/4435 G06F8/443

    摘要: A method to eliminate redundant store and load instruction pairs in the optimization of a stack-based language comprises determining data dependencies within a sequence of instructions, identifying a store-load instruction pair within the sequence of instructions, identifying one or more stack location insensitive sequences between the store-load instruction pair that enclose one or more of the data dependencies, reordering the one or more stack location insensitive sequences based on the data dependencies to place the two instructions of the store-load instruction pair immediately adjacent to each other, and removing the store-load instruction pair.

    摘要翻译: 在优化基于堆栈的语言中消除冗余存储和加载指令对的方法包括确定指令序列内的数据依赖性,识别指令序列内的存储器加载指令对,识别一个或多个堆栈位置不敏感序列 在包含一个或多个数据依赖性的存储 - 加载指令对之间,基于数据依赖性重新排序一个或多个堆栈位置不敏感序列,以将存储加载指令对的两个指令彼此紧邻相邻,以及 删除存储加载指令对。

    Methods and apparatus to implement annotation based thunking
    4.
    发明授权
    Methods and apparatus to implement annotation based thunking 失效
    实施基于注解的方法和装置

    公开(公告)号:US07596783B2

    公开(公告)日:2009-09-29

    申请号:US11440850

    申请日:2006-05-25

    IPC分类号: G06F9/44

    CPC分类号: G06F9/455

    摘要: Methods and apparatus to implement annotation based thunking are disclosed. An example method comprises locating a parameter of a function, the parameter to be passed as a pointer if a size of the parameter is greater than a threshold and to be passed as data if the size of the parameter is not greater than the threshold, and adding an annotation record for the parameter to a byte code image file containing byte code for the function.

    摘要翻译: 披露了实现基于注解的拆分的方法和装置。 一种示例性方法包括:如果参数的大小大于阈值,则定位要作为指针传递的参数的参数,并且如果参数的大小不大于阈值则作为数据传递;以及 将该参数的注释记录添加到包含该函数的字节码的字节码图像文件中。