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
    • 1. 发明公开
    • 정보 처리 장치, 연산 장치 및 정보 전송 방법
    • 信息处理设备,算术设备和信息传输方法
    • KR1020130111167A
    • 2013-10-10
    • KR1020120108256
    • 2012-09-27
    • 후지쯔 가부시끼가이샤
    • 오카다세이시우에키도시카즈고이누마히데유키
    • G06F15/163
    • G06F9/546
    • PURPOSE: A data processing unit, an operation unit, and a data transfer method are provided to, in case a node malfunctions during inter-processing communication where a multi-node system technology is applied, prevent the malfunction from spreading to other nodes. CONSTITUTION: A memory unit remembers a pointer set, which comprises a write pointer indicating the address of storing data transmitted or received among nodes into a memory device; and a read pointer indicating the address of reading the data from the memory device. A notification unit informs a transmitter node of a pointer identifier representing the point set. When data from the transmitter node and the pointer identifier notified from the notification node to the transmitter node are received, a storage unit stores the data in the memory device at the address indicated by the write pointer in the pointer set represented by the pointer identifier. [Reference numerals] (30,30a,30c) Hardware; (33,33a,33c) Driver; (AA) Transmission process #A; (BB) Transmission message #A; (CC) Transmission process #B; (DD) Transmission message #B; (EE) Reception process; (FF) Falling monitoring or interrupt detection; (GG) Reception message; (HH,JJ) Transmission register; (II) Message #A; (KK) Message #B; (LL) Register set #n; (MM) Lead pointer #n; (NN) Right pointer #n; (OO) Reception buffer; (PP) Message #A; (QQ) Message #B; (RR) Shared memory or local memory; (SS) Inter-node interconnect
    • 目的:提供数据处理单元,操作单元和数据传输方法,以便在应用多节点系统技术的处理间通信期间节点发生故障的情况下,防止故障扩散到其他节点。 构成:存储器单元记住指针集,其包括指示将节点之间发送或接收的数据存储到存储器件中的地址的写指针; 以及指示从存储器件读取数据的地址的读指针。 通知单元通知发射机节点表示该点集的指示标识符。 当接收到来自发送器节点的数据和从通知节点通知给发送器节点的指针标识符时,存储单元将数据存储在由指针标识符表示的指针集中由写指针指示的地址处。 (30,30a,30c)硬件; (33,33a,33c)司机; (AA)传输过程#A; (BB)传输消息#A; (CC)传输过程#B; (DD)传输消息#B; (EE)接待流程; (FF)下降监视或中断检测; (GG)接收信息; (HH,JJ)传输寄存器; (二)消息#A; (KK)消息#B; (LL)寄存器集#n; (MM)引导指针#n; (NN)右指针#n; (OO)接收缓冲区; (PP)Message #A; (QQ)消息#B; (RR)共享内存或本地内存; (SS)节点间互连
    • 3. 发明公开
    • 스누프 태그의 제어 장치
    • SNOOP标签控制装置
    • KR1020080016437A
    • 2008-02-21
    • KR1020070049575
    • 2007-05-22
    • 후지쯔 가부시끼가이샤
    • 하타이다마코토우에키도시카즈이시즈카다카하루야마모토다카시호소카와유카오와키다케시이토다이스케
    • G06F9/06
    • G06F12/0822G06F12/123
    • A device for controlling snoop tags is provided to require no index lock, process a read request of the same index promptly, and improve latency by entering the index to a queue and suspending the process of a replace request when the replace request is received. A snoop tag storing part(6) stores the snoop tags corresponding to contents of cache memories of a processor. A queue(2) stores way information and the index of a replace target as an entry for the replace request received from the processor. A comparator(3) compares the index stored in the queue with the index of the read request. A process based on the read request matched with the index is performed for the snoop tag storing part depending on a comparison result. The replace request is invalidated without storing the replace request to the queue if space is not found in the queue when the replace request is received. A replacing part(4) replaces the way information, which is the replace target of the read request matched with the index, with the way information of the entry matched with the index in the queue.
    • 提供了一种用于控制窥探标签的设备,以便不需要索引锁定,及时处理相同索引的读取请求,并通过在接收到替换请求时将索引输入队列并暂停替换请求的过程来改善延迟。 窥探标签存储部分(6)存储与处理器的高速缓冲存储器的内容相对应的窥探标签。 队列(2)将路由信息和替换目标的索引作为从处理器接收的替换请求的条目存储。 比较器(3)将存储在队列中的索引与读请求的索引进行比较。 根据比较结果,针对窥探标签存储部分执行基于与索引匹配的读取请求的处理。 如果在接收到替换请求时没有在队列中找到空格,则替换请求无效,而不将替换请求存储到队列中。 替换部分(4)将与索引匹配的读取请求的替换目标的方式信息替换为与队列中的索引匹配的条目的路径信息。
    • 5. 发明公开
    • 반도체 집적 회로 및 그 제어 방법, 및 정보 처리 장치
    • 半导体集成电路及其控制方法及信息处理装置
    • KR1020090124934A
    • 2009-12-03
    • KR1020090041976
    • 2009-05-14
    • 후지쯔 가부시끼가이샤
    • 야마모토다카시이시즈카다카하루우에키도시카즈오와키다케시모로사와아츠시
    • G11C7/00G11C11/413
    • G01R31/31708G01R31/3004G01R31/31721G06F11/22
    • PURPOSE: A semiconductor IC, a control method thereof, and an information processing apparatus are provided to interpret a data pattern, power noise, and correlation between the data pattern and power noise in a data error state by detecting data of a bus and a noise amount before/after a detection timing of an error. CONSTITUTION: A circuit block(210) is connected to an operation processing unit through a bus. Power noise data signal generation units(220,230,231,232) converts an analog signal of power noise, generated from power for operating the circuit block, into a digital signal. The power noise data signal generation units generate a power noise data signal. An error detecting unit(260) detects an error of data outputted from the circuit block to the bus. A write control unit(250) controls an operation of writing data on the bus and power noise information, based on the power noise data signal, in a recording unit(280). According to a detection timing of the error, the write control unit stops the writing operation.
    • 目的:提供半导体IC及其控制方法以及信息处理装置,通过检测总线的数据和噪声来解释数据模式,功率噪声以及数据模式与功率噪声之间的相关性 在错误的检测定时之前/之后的量。 构成:电路块(210)通过总线连接到操作处理单元。 功率噪声数据信号发生单元(220,230,231,232)将由用于操作电路块的功率产生的功率噪声的模拟信号转换为数字信号。 功率噪声数据信号产生单元产生功率噪声数据信号。 误差检测单元(260)检测从电路块输出到总线的数据的错误。 写入控制单元(250)控制在记录单元(280)中基于功率噪声数据信号在总线上写入数据和功率噪声信息的操作。 根据错误的检测定时,写入控制单元停止写入操作。
    • 6. 发明公开
    • 멀티프로세서 시스템
    • 多处理器系统
    • KR1020080016422A
    • 2008-02-21
    • KR1020070020134
    • 2007-02-28
    • 후지쯔 가부시끼가이샤
    • 우에키도시카즈이시즈카다카하루하타이다마코토야마모토다카시호소이유카오와키,다케시이토우,다이스케
    • G06F15/16G06F9/46
    • G06F12/0813G06F12/1072
    • A multiprocessor system is provided to reduce memory read latency in a local node by reducing data/address latency occurring in memory read which is executed in each system board forming an SMP(Symmetric Multiple Processors) structure assigning tasks to all CPUs symmetrically. A determiner determines whether a read command issued from a CPU(1) and inputted to a global address crossbar(8) is the read command for a memory(2) installed in the system board. An executor executes the read command before global access based on the address notified from the global address crossbar when the inputted read command is the read command for the internal memory. A setting part sets the data read from the memory to be queued to a data queue formed in a CPU side without queuing the read data to the data queue formed in a memory side. An orderer orders the data queue formed in the CPU side to discard or transmit the data to the CPU based on notification received from the global address crossbar.
    • 提供多处理器系统以通过减少在每个系统板中执行的存储器读取中发生的数据/地址等待时间来减少本地节点中的存储器读延迟,形成对称地向所有CPU分配任务的SMP(对称多处理器)结构。 确定器确定从CPU(1)发出并输入到全局地址交叉开关(8)的读取命令是否是安装在系统板中的存储器(2)的读取命令。 当输入的读取命令是内部存储器的读取命令时,执行器根据从全局地址横杠通知的地址在全局访问之前执行读取命令。 设置部分将从待排队的存储器读取的数据设置为在CPU侧形成的数据队列,而不将读取的数据排队到形成在存储器侧的数据队列。 订货人根据从全球地址交叉栏收到的通知命令CPU侧形成的数据队列将数据丢弃或发送到CPU。
    • 10. 发明公开
    • 멀티프로세서 시스템, 시스템 보드 및 캐시 대체 요구 처리방법
    • 多处理器系统,系统板和缓存处理请求处理方法
    • KR1020080016419A
    • 2008-02-21
    • KR1020070018303
    • 2007-02-23
    • 후지쯔 가부시끼가이샤
    • 이시즈카다카하루우에키도시카즈하타이다마코토야마모토다카시호소카와유카오와키다케시이토다이스케
    • G06F15/177G06F15/16
    • G06F12/0833
    • A multiprocessor system, a system board, and a method for handling a cache replacement request are provided to handle the cache replacement request efficiently by preventing load on a global bus and generation of unnecessary eviction caused from the cache replacement request. A multiprocessor system includes a plurality of system boards(100a-100d) respectively having CPU(120a,120b) and a request handler(114) handling the request issued from the CPU, and an address crossbar board(200) arbitrating the request of each system board. The system board includes a cache replacement request loopback circuit(140) determining whether the request issued from the internal CPU. The cache replacement request loopback circuit transfers the request to the address crossbar board when the determined request is not the cache replacement request, and transfers the request to the internal request handler without transferring the request to the address crossbar board when the determined request is the cache replacement request. The cache replacement request loopback circuit includes a loopback queue keeping the cache replacement request issued from the internal CPU.
    • 提供了一种多处理器系统,系统板和用于处理高速缓存替换请求的方法,以通过防止全局总线上的负载并产生由高速缓存替换请求引起的不必要的驱逐来有效地处理高速缓存替换请求。 多处理器系统包括分别具有CPU(120a,120b)和处理从CPU发出的请求的请求处理器(114)的多个系统板(100a-100d)和地址横栏(200)仲裁每个 系统板。 系统板包括高速缓存替换请求回送电路(140),用于确定从内部CPU发出的请求。 当确定的请求不是高速缓存替换请求时,高速缓存替换请求回送电路将请求传送到地址交叉开关板,并且当确定的请求是高速缓存时,将请求传送到内部请求处理程序而不将请求传送到地址横栏 更换请求。 高速缓存替换请求环回电路包括保持从内部CPU发出的高速缓存替换请求的环回队列。