Notice
Recent Posts
Recent Comments
Link
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 |
Tags
- cache optimization
- pipline hazards
- branch prediction
- multibanekd cache
- theta join
- 관계형 모델
- sql
- atomic exchange
- ISA
- relational model
- speculative execution
- register renaming
- moesi
- transactional memory
- cache coherence miss
- pipelined
- nonblocking cache
- Subquery
- sequential consistency
- dependence
- pipelined cache
- load linked
- directory based coherence protocol
- structural hazard
- Cache
- mesi
- cache coherence
- way prediction
- store conditional
- dynamic scheduling
Archives
- Today
- Total
목록DSA (1)
공대생의 공부흔적
[컴퓨터구조#17] Domain Specific Architecture - TPU
이번 글에서는 domain specific 아키텍처에 대해 알아보자. 목차DSA 소개예시 - TPU1. DSA 소개무어의 법칙(칩 내 트랜지스터 점점 더 많아짐)은 CPU, GPU에서 다음과 같은 일반적인 개선사항들을 가능하게 했다.깊은 메모리 위계, 넓은 SIMD unit, 깊은 파이프라인, 브랜치 예측, OoO 실행, speculative prefetching, Multithreading, Multiprocessing이는 아키텍처를 인식하지 않는 소프트웨어에서 성능을 뽑아내는 데 목표가 있었다.이와 달리 DSA는 domain에 specific하게 설계된 아키텍처이다. 최근에는 거의 머신러닝을 타겟으로 하지만, 주요 연산 커널을 정의할 수 있다면 어떤 응용이든 가능하다. DSA를 위한 몇 가지 가이드라..
Computer Architecture
2024. 6. 10. 14:30