发明名称
摘要 PURPOSE:To process a branch instruction with a higher speed by quickening a fetching operation for a branch destination instruction word of a branch instruction in a pipe line controlling system in a large computer. CONSTITUTION:In a case when a load instruction is stored in an instruction register 3 and an unconditional instruction in a branch register 4, a controlling part 12 stores a virtual address computed from an address adder 6 in a virtual address register 8, and fetches an instruction word according to a value of the register 8 against a cash memory 9 and stores in an instruction buffer 1. Since in this register 7 an operand address has been computed according to an instruction of the register 3, if an unconditional branch instruction is detected in the register 4, then a branch prefetching operation is preferred, thereby allowing an operand fetching operation to be done after a branch destination instruction word fetching operation.
申请公布号 JPS6047618(B2) 申请公布日期 1985.10.22
申请号 JP19800108852 申请日期 1980.08.08
申请人 NIPPON ELECTRIC CO 发明人 KONDO TADAO
分类号 G06F9/38 主分类号 G06F9/38
代理机构 代理人
主权项
地址