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
    • 33. 发明专利
    • TOD COUNTING SIGNAL GENERATING CIRCUIT
    • JPH02294713A
    • 1990-12-05
    • JP11494589
    • 1989-05-10
    • HITACHI LTD
    • YAMAGATA MAKOTOISHIKAWA SUKETAKA
    • G06F1/14
    • PURPOSE:To obtain a group of counters whose updating time is completely the same by mutually transferring a synchronizing signal synchronized by a synchronizing circuit between its own system and the other system, and selecting and outputting a signal outputted first out of synchronizing signals from the self-system and the other system as a synchronizing signal for updating a TOD counter. CONSTITUTION:Both the systems are respectively provided with synchronizing circuits 12, 22 to be driven by a common clock 1, and time-of-day (DOD) count signal forming circuits 10, 20 which synchronizes synchronizing signals with nearly constant period sent from the other and its own systems and generates a synchronizing signal, are respectively provided with synchronizing signal sending circuits. The synchronizing signal sending circuits mutually send synchronized synchronizing signals and each of them selects and outputs a synchronized signal output first out of the synchronizing signals in the its own system and the other system as the updating signal of the TOD counter and updates the TOD counter by the updating signal. Consequently, even when a circuit has plural synchronizing circuits, their updating time can be allowed to coincide and all TOD values can always be allowed to coincide with each other.
    • 35. 发明专利
    • CHANNEL CONTROLLER
    • JPS63208148A
    • 1988-08-29
    • JP4021187
    • 1987-02-25
    • HITACHI LTD
    • NINOMIYA KAZUHIKOYAMAGATA MAKOTO
    • G06F13/12
    • PURPOSE:To improve the independence of a channel device by providing plural data buffer registers, all allocating selectively the data buffer register of an unused state, when a memory access request from the channel device has been received. CONSTITUTION:When a channel device 100 executes an input/output operation by a write command, the channel device 100 requests data transfer to a channel controller 200. The channel controller 200 receives data transfer request from the channel 100, when one of buffer registers 220-222 is in an unused state. At the time of requesting data transfer from the device 100, when all of the registers 220-222 are in an unused state, the channel controller 200 allocates the data buffer register 220 to the channel 100, as a use state by a data transfer request by the channel 100, and requests data fetching to a memory device 300. When a fetching operation is completed, the memory device 300 loads fetch data onto a bus 320, and also, informs a memory access completing signal to the device 200.
    • 37. 发明专利
    • PROCESSOR
    • JP2003114802A
    • 2003-04-18
    • JP2001308255
    • 2001-10-04
    • HITACHI LTD
    • IMORI HIROMITSUYAMAGATA MAKOTO
    • G06F9/38G06F9/46G06F9/48G06F11/30
    • PROBLEM TO BE SOLVED: To prevent unnecessary system failure and enhance reliability of the system. SOLUTION: Assuming that a unit 101 requests a unit 102 to execute a store instruction and requests a unit 103 to execute a rfid instruction for switching from an OS to a user program (UP), the unit 103 causes a register 106 to have an identification bit 114 of '1' (UP) and notifies a circuit 107 of the execution of the rfid instruction. The circuit 107 sets a flag 108 and sets a register 109 to '0'. The value of the register 109 is added by a +1 circuit 110, and compared with the value of a register 112 by an identity circuit 111. When the values are identical, the circuit 107 resets the flag 108 to '0'. A failure notice is given from the unit 102 to the unit 104 and the circuit 107, and thereby, when the flat 108 is '1', the circuit 107 sets a flag 113 to '1' as failure during surveillance. Otherwise, the flag 113 is set to '0'. When the flag 113 is '1', the failure is handled as occurring in the OS. When the flag 113 is '0', the failure is handled as occurring in the UP.
    • 40. 发明专利
    • FAULT PROCESSING SYSTEM FOR INSTRUCTION PROCESSOR
    • JP2000181794A
    • 2000-06-30
    • JP36029398
    • 1998-12-18
    • HITACHI LTD
    • JIN KENJIYAMAGATA MAKOTOIMORI HIROMITSU
    • G06F12/08G06F12/16
    • PROBLEM TO BE SOLVED: To make an intermittent/fixed judgement and to secure the reliability while maintaining the performance of the whole device without disconnecting a memory in case of an intermediate fault by handling a fault as a fixed fault only when a fault frequency reaches a certain value. SOLUTION: The fault caused in a WAY0 is reported to a fault frequency counter 10 and a memory control circuit 9 since no retrieval data sequence is present in a WAY1. At this time, the count of the fault frequency counter 10 is set to one. Then data are rewritten and then a parity check is made; if a fault is reported even at this time, the fault frequency counter 10 further counts up by one and an indication for data rewriting is sent again from the memory control circuit 9. If data rewriting and a parity check are carried out four times and a fault is detected in a 5th check, a fixed fault is judged here for the 1st time and the memory is disconnected. Consequently, the reliability of the whole device can be maintained by slightly increasing the circuit scale.