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
    • 2. 发明授权
    • Request processing system, method and program product
    • 请求处理系统,方法和程序产品
    • US09489235B2
    • 2016-11-08
    • US13104710
    • 2011-05-10
    • Takeshi Ogasawara
    • Takeshi Ogasawara
    • G06F7/00G06F17/30G06F9/48G06F12/02
    • G06F9/4881G06F12/0269G06F17/30138G06F2209/485G06F2209/548
    • A processing method has been claimed for reducing the average wait time of requests in a queue in a system environment where garbage collection may occur. In the method, a computer system treats as a unit each request in a queue and a completion time of garbage collection that may occur at the time of processing the request, and processes requests preferentially and systematically in ascending order of the processing times of the units including the garbage collection times, thereby, reducing the average wait time of the requests. While, the computer system managing the queue knows the remaining amount of heap just before processing a certain request, the computer system statistically calculates in advance the amounts of heap to be consumed on a request type basis and holds the values. Accordingly, before processing a certain request, the computer system can predict whether or not the processing of the request will cause garbage collection, in consideration of the estimated heap consumption of the request and the remaining amount of heap in the memory.
    • 已经要求一种处理方法来减少在可能发生垃圾回收的系统环境中的队列中的请求的平均等待时间。 在该方法中,计算机系统将处理请求时可能发生的队列中的每个请求和垃圾收集的完成时间视为单位,并且按照单元的处理时间的升序优先和系统地处理请求 包括垃圾收集时间,从而减少请求的平均等待时间。 管理队列的计算机系统在处理特定请求之前知道剩余的堆量,计算机系统根据请求类型来预先统计计算要消耗的堆的数量并保存该值。 因此,在处理某个请求之前,计算机系统可以考虑到请求的估计堆消耗和存储器中堆的剩余量,来预测请求的处理是否将导致垃圾回收。
    • 5. 发明授权
    • Object optimal allocation device, method and program
    • 对象最优分配设备,方法和程序
    • US09009715B2
    • 2015-04-14
    • US12898845
    • 2010-10-06
    • Takeshi Ogasawara
    • Takeshi Ogasawara
    • G06F9/46G06F9/52
    • G06F9/526
    • A method, system and computer program product for optimally allocating objects in a virtual machine environment implemented on a NUMA computer system. The method includes: obtaining a node identifier; storing the node identifier in a thread; obtaining an object identifier of a lock-target object from a lock thread; writing a lock node identifier into the lock-target object; traversing an object reference graph where the object reference graph contains an object as a graph node, a reference from the first object to a second object as an edge, and a stack allocated to a thread as the root node; determining whether a move-target object contains the lock node identifier; moving the move-target object to a subarea allocated to a lock node if it contains the lock node identifier, and moving the move-target object to the destination of the current traversal target object if the lock node identifier is not found.
    • 一种用于在NUMA计算机系统上实现的虚拟机环境中最佳分配对象的方法,系统和计算机程序产品。 该方法包括:获得节点标识符; 将节点标识符存储在线程中; 从锁定线程获取锁定目标对象的对象标识; 将锁定节点标识符写入锁定目标对象; 遍历对象引用图,其中对象引用图包含对象作为图形节点,从第一对象到第二对象作为边缘的引用,以及分配给作为根节点的线程的堆栈; 确定移动目标对象是否包含所述锁定节点标识符; 将移动目标对象移动到分配给锁定节点的子区域(如果其包含锁定节点标识符),并且如果未找到锁定节点标识符,则将移动目标对象移动到当前遍历目标对象的目的地。
    • 7. 发明授权
    • Method and system to accelerate address translation
    • 加快地址转换的方法和系统
    • US08806146B2
    • 2014-08-12
    • US12970046
    • 2010-12-16
    • Megumi ItoTakeshi Ogasawara
    • Megumi ItoTakeshi Ogasawara
    • G06F12/00
    • G06F12/1009G06F12/1475G06F2212/652
    • In a method to accelerate address translation into a physical address, a computer maps a virtual memory area with a large page, the virtual memory area including multiple virtual pages satisfying a predetermined condition and being handled in units of pages, the large page having a larger area than each of the virtual pages, and under a condition in which one of the virtual pages mapped with and included in the large page has a memory protection attribute different from a memory protection attribute of the other virtual page, sets physical memory protection information for protecting a physical page corresponding to the one virtual page having the different memory protection attribute.
    • 在将地址转换加速到物理地址的方法中,计算机将虚拟存储器区域与大页面进行映射,虚拟存储器区域包括满足预定条件的多个虚拟页面并以页面为单位处理,大页面具有较大的页面 区域,并且在与大页面映射并包含在大页面中的一个虚拟页面具有与另一个虚拟页面的存储器保护属性不同的存储器保护属性的条件下,设置用于 保护对应于具有不同存储器保护属性的一个虚拟页面的物理页面。