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
    • 73. 发明专利
    • METHOD FOR ACCELERATING OPERATING SPEED OF PROCESSOR
    • JPH05143335A
    • 1993-06-11
    • JP13867091
    • 1991-05-15
    • SUN MICROSYSTEMS INC
    • ERITSUKU HAATOUIGU JIENSEN
    • G06F9/312G06F9/34G06F9/38
    • PURPOSE: To improve an operating speed by deciding the difference of the effective addresses of two load instructions indexed by the address of the load instruction, and adding the difference in order to obtain the predictive address of the next load instruction. CONSTITUTION: Each time an instruction for executing a predictive operation reaches a decoding stage, a predictive table 12 is indexed by using the address of the instruction in order to retrieve the entry of an LEA and an NLEA. The tag of a line to which access is performed is compared with the tag of the instruction, and when an upper bit is the same and a state bit is 1, the values of the effective addresses of the LEA string and the NLEA string are retrieved, and transferred to a subtracter. The subtracter supplies the difference of two sets of lower bits as an output value to an adder with the front effective address. The predicted effective address is compared with the effective address of a load instruction, an enable output is supplied to an AND gate, and information in the predictive address can be used by an instruction following the load instruction.
    • 74. 发明专利
    • DATA PROCESSOR
    • JPH04260930A
    • 1992-09-16
    • JP523091
    • 1991-01-21
    • MITSUBISHI ELECTRIC CORP
    • YOSHIDA TOYOHIKO
    • G06F9/38G06F9/30G06F9/312G06F12/00
    • PURPOSE:To offer a data processor capable of loading plural data from a memory to a register by one processing and highly efficiently processing an instruction for storing data from the register to the memory with respect to a data processor for decomposing an instruction for transferring plural data between register files and the memory into plural transfer instructions by a decoder and executing the pipeline processing of the decomposed instructions. CONSTITUTION:The data processor is provided with an instruction decoding part 112 for decomposing both of an instruction for loading plural data from the memory to the register and an instruction for storing plural data from the register in the memory to plural processing units and decoding respective processing units, an instruction execution part connected to an external memory through a data bus to execute instructions and an instruction execution control part for controlling the execution of instructions by the instruction execution part in accordance with respective processing units outputted from the decoding part 112 and the decoding processing of the decoding part 112 to output respective processing units and the instruction execution of the instruction execution part based upon the outputted processing units are mutually overlapped and executed in accordance with the principle of pipeline processing.
    • 79. 发明专利
    • Controller of electronic equipment
    • 电子设备控制器
    • JPS5955538A
    • 1984-03-30
    • JP16508982
    • 1982-09-24
    • Casio Comput Co Ltd
    • IBA AKIO
    • G06F9/26G06F9/22G06F9/312G06F9/32G10H7/00
    • G06F9/30043
    • PURPOSE:To simplify the constitution of a memory, by providing an address means consisting of an adder circuit and a latch circuit and a selection means consisting of plural gate circuits, to read out and process the data after setting it to a program region. CONSTITUTION:When a data storing instruction is supplied, a control signal generating part CONT delivers a pseudo instruction signal. Then an AND3 is turned on and an FF1 is set to deliver an ''NOP'' signal. The address data of an Rg3 existing on an address line is stored in an RAM through a Gg2 which is turned on by the output GOST of an NOR1 and via half adders HA2 and an Rg10. When a Gg5 is turned on by the output of an AND7, the data of 4 bits given from the Rg3 is fed to the RAM through a bus line. Here an R/W signal is delivered from the CONT to write the 4-bit data to an address designated by an address input ADD. Thereafter, the writing is successively carried out in the same way.
    • 目的:为了简化存储器的结构,通过提供由加法器电路和锁存电路组成的地址装置和由多个门电路组成的选择装置,在将其设置到程序区域之后读出并处理数据。 构成:当提供数据存储指令时,控制信号生成部CONT发送伪指令信号。 然后打开一个AND3,一个FF1被设置为提供一个“NOP”信号。 存在于地址线上的Rg3的地址数据通过由NOR1的输出GOST和半加法器HA2和Rg10导通的Gg2存储在RAM中。 当通过AND7的输出打开Gg5时,从Rg3给出的4位的数据通过总线被馈送到RAM。 这里,从CONT发送R / W信号,以将4位数据写入由地址输入ADD指定的地址。 此后,以相同的方式依次进行写入。
    • 80. 发明专利
    • Mark signal generating circuit
    • 标记信号发生电路
    • JPS58213323A
    • 1983-12-12
    • JP9482382
    • 1982-06-04
    • Hitachi Ltd
    • MORIKAWA TAKASHI
    • G06F13/12G06F9/312
    • G06F9/30043
    • PURPOSE:To save the amount of hardware, by providing two pointer effectiveness displays for discriminating whether a start pointer and an end pointer are used or not. CONSTITUTION:A mark signal generating circuit is provided with a start pointer effectiveness indicating flip-flop 7, an end pointer effectiveness displaying flip- flop 8, an AND circuit 9 and an OR circuit 10. A value of a start pointer 1 is inputted to an A pattern generator 3 as it is from the AND circuit 9 when the flip-flop 7 is 1, and a value of all 0 is inputted as the start pointer when the flip-flop 7 is 0. Further, a value of an end pointer 2 is inputted to a B pattern generator 4 as it is when the flip-flop 8 is 1, and all 1 is inputted as the end pointer when 0. Thus, two pointers are not required at the memory transfer on the halfway.
    • 目的:为了节省硬件量,通过提供两个指针有效性显示来区分是否使用起始指针和结束指针。 构成:标记信号发生电路具有指示触发器7的起始指针效能,结束指针有效性显示触发器8,AND电路9和OR电路10.起始指针1的值被输入到 当触发器7为1时,来自AND电路9的A模式发生器3,并且当触发器7为0时,输入全0的值作为开始指针。此外,结束的值 当触发器8为1时,指针2被直接输入到B模式发生器4,并且当0时,指针2作为结束指针输入。因此,在中途的存储器传送中不需要两个指针。