|
|
- # 理解页置换算法
- ## 选择四种替换算法（0：LRU置换算法，1:改进的clock 页置换算法，2：工作集页置换算法，3：缺页率置换算法）中的一种来设计一个应用程序（可基于python, ruby, C, C++，LISP等）模拟实现，并给出测试。请参考page-replacement-policy.py 代码或独自实现。
-
- ###page-replacement-policy.py 代码
-
- #### 对模拟环境的抽象
-
- 虚拟页访问序列：addresses=1,2,3,4,0,5.... 这里面的最大值代表最大页号
- 页替换算法：policy=FIFO, LRU, OPT, CLOCK
- CLOCK算法用的bit位数：clockbits=1
- 物理页帧大小：cachesize
- 实际保持的也访问序列：addrList [1,2,3,4,0,5,...]
- 物理页帧内容：memory [] 初始为空
- 当前占用的页帧数量：count 初始位0
-
- #### 执行过程描述
-
-
- ```
- for nStr in addrList:
- # for clock need to track the ref by reference bits
-
- try:
- idx = memory.index(n)
- hits = hits + 1
- if policy == 'LRU' :
- ....
- except:
- idx = -1 #missing
- miss = miss + 1
-
- if idx=-1 and ...: #missing and need replacement
- #if phy page frames are full
- # for FIFO , LRU
- # replace victim item from memory by " victim = memory.pop(0)"
-
- # for CLOCK
- # find one page for the beginning of scan
- # check ref[page] and update ref[page]
- # find a victim which ref[page]=0 by memory.remove(page)
- else:
- # miss, but no replacement needed (phy page frame not full)
- # now add to memory
-
- #update ref for clock replacement
- ```
|
