This shows you the differences between two versions of the page.
Both sides previous revision Previous revision | Next revision Both sides next revision | ||
buzzword [2014/01/17 19:20] rachata |
buzzword [2014/01/22 19:23] rachata |
||
---|---|---|---|
Line 164: | Line 164: | ||
* Optimization opportunity | * Optimization opportunity | ||
+ | ===== Lecture 4 (1/22 Wed.) ===== | ||
+ | |||
+ | * Semantic gap | ||
+ | * Small vs. Large semantic gap (CISC vs. RISC) | ||
+ | * Benefit of RISC vs. CISC | ||
+ | * Micro operations/microcode | ||
+ | * Translate complex instructions into smaller instructions | ||
+ | * Parallelism (motivation for RISC) | ||
+ | * Compiler optimization | ||
+ | * Code optimization through translation | ||
+ | * VLIW | ||
+ | * Fixed vs. variable length instructions | ||
+ | * Tradeoffs | ||
+ | * Alignment issues? (fetch/decode) | ||
+ | * Decoding issues? | ||
+ | * Code size? | ||
+ | * Adding additional instructions? | ||
+ | * Memory bandwidth and cache utilization? | ||
+ | * Energy? | ||
+ | * Encoding in variable length instructions | ||
+ | * Structure of Alpha instructions and other uniform decode instructions | ||
+ | * Different type of instructions | ||
+ | * Benefit of knowing what type of instructions | ||
+ | * Speculatively operate future instructions | ||
+ | * x86 and other non-uniform decode instructions | ||
+ | * Tradeoff vs. uniform decode | ||
+ | * Tradeoffs for different number of registers | ||
+ | * Spilling into memory if the number of registers is small | ||
+ | * Compiler optimization on how to manage which value to keep/spill | ||
+ | * Addressing modes | ||
+ | * Benefits? | ||
+ | * Types? | ||
+ | * Different uses of addressing modes? | ||
+ | * Various ISA-level tradeoffs | ||
+ | * Virtual memory | ||
+ | * Unalign memory access/aligned memory access | ||
+ | * Cost vs. benefit of unaligned access | ||
+ | * ISA specification | ||
+ | * Things you have to obey/specifie in the ISA specification | ||
+ | * Architectural states | ||
+ | * Microarchitecture implements how arch. state A transformed to the next arch. state A' | ||
+ | * Single cycle machines | ||
+ | * Critical path in the single cycle machine | ||
+ | * Multi cycle machines | ||
+ | * Functional units | ||
+ | * Performance metrics | ||
+ | * CPI/IPC | ||
+ | * CPI of a single cycle microarchitecture |