发明名称 ZERO CYCLE LOAD
摘要 The present invention includes a system and a method for reducing the latency of load operations. The load command language in which a register renaming unit within a processor is decoded to determine the conversion of the cycle load operation with the agent whether it is fit for the conversion or not. The physical register identifier associated with the source operand of the dependence store instruction and the control logic assigned to the destination location operand of the load command language are allocated. Moreover, a register renaming unit marks the load command language to prevent the load command language reading data associated with the source operand of the store instruction to be decoded from the memory. Due to the reproduction of the renaming, this data can be transmitted to newer instructions which are more dependent on the load command language from the physical register file. [Reference numerals] (100) Computer system;(110) Microprocessor;(112) Processor core;(114) Physical registers;(116) Cash memory subsystem;(118) Interface logic;(120) Memory controller;(122) Memory bus;(130) Disk memory;(150a,150b) Peripherals
申请公布号 KR20130140582(A) 申请公布日期 2013.12.24
申请号 KR20130068008 申请日期 2013.06.13
申请人 APPLE INC. 发明人 WILLIAMS III GERARD R.;MYLIUS JOHN H.;BLASCO ALLUE CONRADO
分类号 G06F9/06;G06F9/30 主分类号 G06F9/06
代理机构 代理人
主权项
地址