9. Lottery Scheduling

Lottery Scheduling

0. Basic Concept

A : 75 tickets (0~74)
B：25 tickets (75~99)

A1 + A2：1000 tickets
B1：10 tickets
注意：要区分一个进程和一个线程的tickets。
如下所示，A和B分别有100tickets，A中有两个需要处理的任务A1和A2，总共有1000tickets，每个任务各分配500tickets，相当于CPU给A分配的100tickets的50tickets。其实就是CPU给一个进程分配了100tickets，该进程给里面两个线程分配了1000tickets。

1. Stride Scheduling

A: 100 tickets
B: 50 tickets
C: 250 tickets
用10000除以每个job的tickets，得到了对应的stride_value。
A： stride_value = 10000 / 100 = 100
B： stride_value = 10000 / 50 = 200
C： stride_value = 10000 / 250 = 40
刚开始每个job的stride_value都是0，所以依次运行A、B、C。
每次运行后都会增加对应job的stride_value。
当所有job都运行一次后再判断stride_value的值的大小，选取最小的值对应的job运行。

2. The Linux Completely Fair Scheduler (CFS)

A simple counting-based technique: virtual runtime

time_slice = sched_latency / n

sched_latency是系统给出的，n 代表job数量，上图中sched_latency = 48ms，n = 4
故每个未完成的job的time_slice = 12 ms
当job C、D完成后，job A、B的 time_slice = sched_latency / n = 48ms / 2 = 24ms
但是有个问题就是如果n很大，那每个job分配的time_slice不就很小，这样每个job运行的不得劲，故引入了min_granularity保证每个job有最小的time_slice。

CFS也做了另外一个改善就是存储结构从链表变成了红黑树。
链表插入一个元素时间复杂度为O(n)，而红黑树只需要O(log)。

Homework (Simulation)

This program, lottery.py, allows you to see how a lottery scheduler
works. See the README for details.
用的学校机房的电脑~

Question & Answer

1. Compute the solutions for simulations with 3 jobs and random seeds of 1, 2, and 3.

2. Now run with two specific jobs: each of length 10, but one (job 0) with just 1 ticket and the other (job 1) with 100 (e.g., -l 10:1,10:100). What happens when the number of tickets is so imbalanced? Will job 0 ever run before job 1 completes? How often? In general, what does such a ticket imbalance do to the behavior of lottery scheduling?

3. When running with two jobs of length 100 and equal ticket allocations of 100 (-l 100:100,100:100), how unfair is the scheduler? Run with some different random seeds to determine the (probabilistic) answer; let unfairness be determined by how much earlier one job finishes than the other.

我觉得不管seed为多少，两个job完成时间都不会差距很大。但是我也只测验了一部分，样本太小，希望大家能自己多测验…

4. How does your answer to the previous question change as the quantum size (-q) gets larger?

可以看出来 the length of time slice (- q)对job完成时间的影响非常大。

5. Can you make a version of the graph that is found in the chapter? What else would be worth exploring? How would the graph look with a stride scheduler?

to be continue…