python实现的EDF(earliest deadline first)算法
Posted gebicungaha
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python实现的EDF(earliest deadline first)算法相关的知识,希望对你有一定的参考价值。
1. 最早截止时间优先EDF(Earliest DeadlineFirst)算法是非常著名的实时调度算法之一。在每一个新的就绪状态,调度器都是从那些已就绪但还没有完全处理完毕的任务中选择最早截止时间的任务,并将执行该任务所需的资源分配给它。在有新任务到来时,调度器必须立即计算EDF,排出新的定序,即正在运行的任务被剥夺,并且按照新任务的截止时间决定是否调度该新任务。如果新任务的最后期限早于被中断的当前任务,就立即处理新任务。按照EDF算法,被中断任务的处理将在稍后继续进行。
2. 该算法的思想是从两个任务中选择截至时间最早的任务,把它暂作为当前处理任务,再判断该任务是否在当前周期内,若不在当前周期内,就让另一任务暂作当前处理任务,若该任务也不在当前周期内,就让CPU空跑到最靠近的下一个截至时间的开始,若有任务在该周期内,就判断该任务的剩余时间是否小于当前截至时间与当前时间的差,若小于,则让该任务运行到结束.否则,就让该任务运行到该周期的截止时间,就立即抢回处理器,再判断紧接着的最早截至时间,并把处理器给它,做法同上,如此反复执行.
1 #!/user/bin/env python 2 # -*- coding:utf-8 -*- 3 4 5 class GetCloseTime: 6 def __init__(self, dictionary): 7 self.dictionary = dictionary 8 9 def greatest_common_divisor(self, _left, _right): 10 return _left if _right == 0 else self.greatest_common_divisor(_right, _left % _right) 11 12 def lowest_common_multiple(self): 13 temp_result = 1 14 for value in self.dictionary.values(): 15 temp_result = value[1] * temp_result / self.greatest_common_divisor(value[1], temp_result) 16 return temp_result 17 18 19 class TaskControlBlock: 20 CURRENT_TIME = 0 21 22 def __init__(self, dictionary, 23 name_list, 24 period_time, 25 central_processing_unit_time, 26 remain_time, 27 current_period): 28 for key in dictionary.keys(): 29 name_list.append(key) 30 period_time.append(dictionary.get(key)[1]) 31 central_processing_unit_time.append(dictionary.get(key)[0]) 32 remain_time.append(dictionary.get(key)[0]) 33 current_period.append(1) 34 35 def get_index_of_min(self, earliest_deadline_task_list, minimum): 36 return earliest_deadline_task_list.index(minimum) 37 38 def get_another_index_of_min(self, earliest_deadline_task_list, minimum): 39 earliest_deadline_task_list[earliest_deadline_task_list.index(minimum)] = 100000 40 return earliest_deadline_task_list.index(min(earliest_deadline_task_list)) 41 42 def is_execute(self, central_processing_unit_time, period_time): 43 temp_list = [a / b for a, b in zip(central_processing_unit_time, period_time)] 44 return sum(temp_list) 45 46 def scheduling(self, name_list, 47 period_time, 48 central_processing_unit_time, 49 remain_time, 50 current_period): 51 if self.is_execute(central_processing_unit_time, period_time) > 1: 52 print("error, scheduling finish!") 53 exit(0) 54 earliest_deadline_task = self.get_index_of_min( 55 [a * b for a, b in zip(period_time, current_period)], 56 min(a * b for a, b in zip(period_time, current_period))) 57 58 if self.CURRENT_TIME < period_time[earliest_deadline_task] * (current_period[earliest_deadline_task] - 1): 59 current_period_p = period_time[earliest_deadline_task] * (current_period[earliest_deadline_task] - 1) 60 temp_list = [a * b for a, b in zip(period_time, current_period)] 61 while self.CURRENT_TIME < period_time[earliest_deadline_task] * 62 (current_period[earliest_deadline_task] - 1): 63 earliest_deadline_task = self.get_another_index_of_min(temp_list, min(temp_list)) 64 65 if remain_time[earliest_deadline_task] <= current_period_p - self.CURRENT_TIME: 66 running_time = remain_time[earliest_deadline_task] 67 else: 68 running_time = current_period_p - self.CURRENT_TIME 69 # current_period_p = period_time[earliest_deadline_task] * current_period[earliest_deadline_task] 70 remain_time[earliest_deadline_task] -= running_time 71 print(name_list[earliest_deadline_task], self.CURRENT_TIME, running_time) 72 self.CURRENT_TIME += running_time 73 if remain_time[earliest_deadline_task] == 0: 74 current_period[earliest_deadline_task] += 1 75 remain_time[earliest_deadline_task] = central_processing_unit_time[earliest_deadline_task] 76 else: 77 current_period_p = period_time[earliest_deadline_task] * current_period[earliest_deadline_task] 78 if remain_time[earliest_deadline_task] <= current_period_p - self.CURRENT_TIME: 79 running_time = remain_time[earliest_deadline_task] 80 else: 81 running_time = current_period_p - self.CURRENT_TIME 82 remain_time[earliest_deadline_task] -= running_time 83 print(name_list[earliest_deadline_task], self.CURRENT_TIME, running_time) 84 self.CURRENT_TIME += running_time 85 if remain_time[earliest_deadline_task] == 0: 86 current_period[earliest_deadline_task] += 1 87 remain_time[earliest_deadline_task] = central_processing_unit_time[earliest_deadline_task] 88 89 90 if __name__ == "__main__": 91 task_dictionary = {"A": [10, 30], 92 "B": [20, 60], 93 "C": [30, 90]} 94 close_time_object = GetCloseTime(task_dictionary) 95 close_time = close_time_object.lowest_common_multiple() 96 97 current_time = 0 98 name_list = [] 99 period_time = [] 100 central_processing_unit_time = [] 101 remain_time = [] 102 current_period = [] 103 tcb = TaskControlBlock(task_dictionary, 104 name_list, 105 period_time, 106 central_processing_unit_time, 107 remain_time, 108 current_period) 109 110 while tcb.CURRENT_TIME < close_time: 111 tcb.scheduling(name_list, 112 period_time, 113 central_processing_unit_time, 114 remain_time, 115 current_period)
以上是关于python实现的EDF(earliest deadline first)算法的主要内容,如果未能解决你的问题,请参考以下文章
kafka auto.offset.reset latest earliest 详解
kafka auto.offset.reset / latest / earliest 详解