计算机系统结构总结_Memory Hierarchy and Cache Optimization

Posted pdev

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了计算机系统结构总结_Memory Hierarchy and Cache Optimization相关的知识,希望对你有一定的参考价值。

Textbook:

《计算机组成与设计——硬件/软件接口》    HI

《计算机体系结构——量化研究方法》         QR


Ch4. Cache Optimization

本章要讨论的问题就是 How to Improve Cache Performance? 

前面讲过 Average memory access time = HitTime + (MissRate * MissPenalty)

那么我们的方向就是Reduce MissRate / HitTime / MissPenalty

 

1. 6 Basic Cache Optimization(PPT P3)

 Reducing hit time

1. Giving Reads Priority over Writes
E.g., Read complete before earlier writes in write buffer  ??

 

2. Avoiding Address Translation during Cache Indexing

Cache中使用虚拟地址,这样就可以同时Access TLB和Cache / Access Cache firstly

 

Reducing Miss Penalty

3. Multilevel Caches

AMAT = Hit TimeL1 + Miss RateL1 x Miss PenaltyL1
Miss PenaltyL1 = Hit TimeL2 + Miss RateL2 x Miss PenaltyL2

原来Miss PenaltyL1要访问内存,很慢。现在多了L2

 

Reducing Miss Rate

4. Larger Block size (Compulsory misses)

...

 

5. Larger Cache size (Capacity misses)

...

 

6. Higher Associativity (Conflict misses)

...

2. 11 Advanced Cache Optimizations (PPT P12)

• Reducing hit time

1. Small and simple caches(QR  P59)

 

 

2. Way prediction

 

 

3. Trace caches

 

 

Increasing cache bandwidth

4. Pipelined caches

 

 

5. Multibanked caches

 

 

6. Nonblocking caches

 

 

Reducing Miss Penalty

7. Critical word first

 

 

8. Merging write buffers

 

 

Reducing Miss Rate

9. Compiler optimizations

 

 

Reducing miss penalty or miss rate via parallelism

10. Hardware prefetching

 

 

11. Compiler prefetching

 

 

3. Main Memory Optimization (PPT P39)

 

 

 

 

...

 

以上是关于计算机系统结构总结_Memory Hierarchy and Cache Optimization的主要内容,如果未能解决你的问题,请参考以下文章

计算机系统结构总结_Multiprocessor

计算机系统结构总结_Instruction Set Architecture

技术向|内存屏障(Memory Barriers)--Runtime Time

Linux内核分析_课程学习总结报告

SA20225377_邱世茂

数据结构学习笔记——KMP算法中的常见计算题目总结