G mmu cache advanced pipelining etc mmu cache

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: owerPC, SPARC, Precision Architecture. SPARC Embedded: ARM, Hitachi SH, Thumb) JUSTIFICATION OF RISC DESIGNS – Program tracing (~1980): only 25% of CISC instructions are used 95% of the time [LOAD [LOAD,STORE,ADD, JUMP] JUMP] CISC may waste processor real estate » use real estate for more important “functions” (e (e.g., MMU, cache, advanced pipelining, etc.) MMU cache advanced etc » Hardwired implementations have better performance than microcoded ones Some material is adapted from D. Culler & D. Patterson (UCB) S. Ziavras Typical RISC in 1980 Typical RISC in 1980’s • • • • • • • • < 100 instructions Fixed instruction formats 3-5 simple addressing modes Most instructions are register-based Memory access done only by load/store instructions Large register file (>32 registers) Hardwired control Higher clock rate Higher MIPS * rate • Lower CPI ** RISC instructions RISC instructions * MIPS: Million Instructions Per Second (a performance metric) ** CPI: clock Cycles Per Instruction Some material is adapted from D. Culler &...
View Full Document

This document was uploaded on 02/09/2014.

Ask a homework question - tutors are online