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
    • 6. 发明公开
    • 스누프 필터에서의 실효를 감소시키기 위한 라인 스와핑 스킴을 위한 방법, 장치, 시스템 및 컴퓨터 판독 가능 기록 매체
    • 线路切换方案减少SNOOP过滤器中的反向失败
    • KR1020080055681A
    • 2008-06-19
    • KR1020070129626
    • 2007-12-13
    • 인텔 코포레이션
    • 친다마니,선다람쳉,카이만드비웰라,말콤파힘,바하플레더러,케이쓰
    • G06F13/16G06F12/08
    • G06F12/0831G06F12/082
    • A line swapping scheme for reducing invalidation in a snoop filter is provided to reduce the invalidations and improve issuance of back invalidations by selecting a better replacement measure without increasing the size of the snoop filter. A data request is received from one of a plurality of processors(702), and a cache entry position is determined particularly based on the data request(704). Data is stored to the cache entry position in a cache corresponding to the processor(706). Occurrence of cache miss is determined. A consistency record corresponding to the data is stored to affinity corresponding to the cache by storing the data to the cache entry position, which is selected in the affinity corresponding to the cache based on a way-hint replacement policy, when the cache entry position is found in the corresponding affinity(708A). The consistency record is stored to the guided cache entry position of the affinity when the cache entry position is not found in the corresponding affinity(708B).
    • 提供用于减少窥探过滤器中的无效的线路交换方案,以减少无效,并通过选择更好的替换度量而不增加窥探过滤器的大小来改善背部无效的发布。 从多个处理器(702)之一接收数据请求,特别地基于数据请求确定高速缓存条目位置(704)。 将数据存储到与处理器(706)相对应的高速缓存中的高速缓存条目位置。 确定缓存未命中的发生。 将与数据对应的一致性记录通过将数据存储到高速缓存条目位置而被存储为与高速缓存相对应的亲和度,该高速缓存条目位置在高速缓存条目位置为 发现在相应的亲和力(708A)。 当在相应的亲和度(708B)中没有找到高速缓存入口位置时,一致性记录被存储到亲和力的被引导高速缓存条目位置。
    • 8. 发明公开
    • 인-메모리, 인-페이지 디렉토리 캐시 일관성 기법
    • IN-MEMORY,IN-PAGE目录高速缓存方案
    • KR1020090073983A
    • 2009-07-03
    • KR1020080136546
    • 2008-12-30
    • 인텔 코포레이션
    • 스테이너,이안카이조지,종-닝티와리,사우라브쳉,카이
    • G06F13/16G06F9/46G06F9/06G06F12/00
    • G06F12/0817
    • An in-memory and an in-page directory cache coherency technique capable of applying the increase of a processor or processor cores are provided to secure a page hit access which becomes the same as the directory cache hit. A memory access request about a cache line which is requested from one processor is received(302). A memory page storing a required cache line is brought from a memory unit. Coherency information related to the required cache line is accessed from the memory unit(304). The memory page comprises a directory line having the consistency information corresponding to the required cache line. According to the consistency information, data related to the required cache line is read(306). The data is returned by the processor(308).
    • 提供了能够应用处理器或处理器核的增加的内存中和页内目录高速缓存一致性技术以保护与目录高速缓存命中相同的页命中访问。 接收关于从一个处理器请求的高速缓存线的存储器访问请求(302)。 从存储器单元提取存储所需高速缓存行的存储器页。 从存储器单元(304)访问与所需高速缓存行有关的一致性信息。 存储器页面包括具有与所需高速缓存行对应的一致性信息的目录行。 根据一致性信息,读取与所需高速缓存行有关的数据(306)。 数据由处理器返回(308)。