일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | |||
5 | 6 | 7 | 8 | 9 | 10 | 11 |
12 | 13 | 14 | 15 | 16 | 17 | 18 |
19 | 20 | 21 | 22 | 23 | 24 | 25 |
26 | 27 | 28 | 29 | 30 | 31 |
- DS
- Pipelining
- Java
- javascript
- MacOS
- CSS
- php
- mysql
- XML
- Linux
- web
- computer
- DoM
- control
- Algorithm
- Class
- DB
- python
- system
- github
- function
- data structure
- instruction
- while
- DATAPATH
- react
- html
- MIPS
- architecture
- for
- Today
- Total
YYYEJI
[MIPS] Performance 본문
Performance는 컴퓨터의 성능을 의미합니다.
우리가 컴퓨터를 디자인을 한다면 비용&성능을 배제할 수 없을 것입니다.
우선 MIPS를 배우는 지금은 시간에 초점을 맞춰서 고민을 해봅시다.
기능은 수행시간에 반비례합니다.
Performance = 1/Execution Time
CPU 시간은 아낼와 같습니다.
CPU Time = CPU Clock Cycles X Clock Cycle Time
CPU Time = CPU Clock Cycles/Clock Rate
(CPU Cycle Time = 1/Clock Rate)
* Clock rate = frequency
프로그램 수행 시간은 프로그램의 instruction의 개수(instruction count)과 비례합니다.
Seconds/program = Machine Instructions/program
Q) Machine instruction의 개수에 영향을 받는가?
A) 받기 때문에 instruction set을 잘 만들어야 합니다.
Q) ISA은 어떤 것에 영향을 받는가?
A) Compiler에 영향을 받습니다.
*** 즉, execution time은 instruction count과 관련이 있습니다. ***
Q) 어떻게 clock period를 줄일 수 있는가?
A) Critical path(최상경로)를 줄입니다.
그렇게 나온 공식이 있습니다.
Seconds/program = Instructions/program X Cycles/instruction X Seconds/Cycle
CPU time을 줄이는데,
3개의 terms이 모두 중요하기 때문에 모두 신경써야 됩니다.
◡̈
'Computer architectures' 카테고리의 다른 글
[MIPS] block, hit, miss (Memory Hierarchy) (0) | 2022.12.03 |
---|---|
[MIPS] Memory hierarchy (locality) (0) | 2022.12.03 |
[MIPS] Pipelining Hazard (0) | 2022.11.28 |
[MIPS] Pipelining Control Signal (0) | 2022.11.27 |
[MIPS] Corrected Datapath란? (0) | 2022.11.27 |