给定大小为N的数组arr []和代表线程数的整数T ,任务是使用多线程查找所有非重复的数组元素。
例子:
Input: arr[] = { 1, 0, 5, 5, 2}, T = 3
Output: 0 1 2
Explanation:
The frequency of 0 in the array arr[] is 1.
The frequency of 1 in the array arr[] is 1.
The frequency of 2 in the array arr[] is 1.
Therefore, the required output is 0 1 2
Input: arr[] = { 1, 1, 5, 5, 2, 4 }, T = 3
Output: 2 4
Explanation:
The frequency of 2 in the array arr[] is 1.
The frequency of 4 in the array arr[] is 1.
Therefore, the required output is 2 4
方法:想法是使用C++中可用的pthread库为并发进程流创建多个线程,并在多线程程序中执行多个操作(pthread create,pthread join,lock等)。请按照以下步骤解决问题:
- 将数组划分为T个子数组,以便每个大小为N / T的子数组将在单个线程中执行。
- 初始化一个映射,例如mp ,以存储每个数组元素的频率。
- 创建一个pthread_mutex_lock,例如lock1 ,以确保所有线程都不会彼此跳闸并破坏Map容器。
- 定义一个函数func()来执行线程的主体。此函数通常称为线程内核,是在线程创建过程中提供的。
- 使用pthread_mutex_lock()锁定当前线程,使其不与其他线程重叠
- 遍历给定范围作为数组arr []中函数func()的参数,并增加遇到的数组元素的频率。
- 使用函数pthread_mutex_unlock()释放当前线程。
- 初始化一个类型为pthread_t的数组,例如thread [] ,用于存储线程。
- 遍历范围[0,T]并通过调用pthread_create()函数创建线程并将其存储在线程中[i]
- 当每个线程执行其各自的任务时, main()函数将需要等待,直到每个线程完成其工作。
- 使用pthread_join()函数等待,直到每个线程完成执行函数func()为止
- 遍历范围[0,T]并为每个线程[i]调用pthread_create()函数
- 最后,遍历map mp并打印仅出现一次的元素。
下面是上述方法的实现:
C++
// C++ program to implement
// the above approach
#include
#include
using namespace std;
// Structure of subarray
// of the array
struct range_info {
// Stores start index
// of the subarray
int start;
// Stores end index
// of the subarray
int end;
// Stores pointer to the
// first array element
int* a;
};
// Declaring map, and mutex for
// thread exclusion(lock)
map mp;
pthread_mutex_t lock1;
// Function performed by every thread to
// count the frequency of array elements
// in current subarray
void* func(void* arg)
{
// Taking a lock so that threads
// do not overlap each other
pthread_mutex_lock(&lock1);
// Initalize range_info
// for each thread
struct range_info* rptr
= (struct range_info*)arg;
// Thread is going through the array
// to check and update the map
for (int i = rptr->start;
i <= rptr->end; i++) {
// Stores iterator to the map
map::iterator it;
// Update it
it = mp.find(rptr->a[i]);
// If rptr->a[i] not found
// in map mp
if (it == mp.end()) {
// Insert rptr->a[i] with
// frequency 1
mp.insert({ rptr->a[i], 1 });
}
else {
// Update frequency of
// current element
it->second++;
}
}
// Thread releases the lock
pthread_mutex_unlock(&lock1);
return NULL;
}
// Function to find the unique
// numbers in the array
void numbers_occuring_once(int arr[],
int N, int T)
{
// Stores all threads
pthread_t threads[T];
// Stores start index of
// first thread
int spos = 0;
// Stores last index
// of the first thread
int epos = spos + (N / T) - 1;
// Traverse each thread
for (int i = 0; i < T; i++) {
// Initalize range_info for
// current thread
struct range_info* rptr
= (struct range_info*)malloc(
sizeof(struct range_info));
// Update start index of
// current subarray
rptr->start = spos;
// Stores end index of
// current subarray
rptr->end = epos;
// Update pointer to the first
// element of the array
rptr->a = arr;
// creating each thread with
// appropriate parameters
int a
= pthread_create(&threads[i], NULL,
func, (void*)(rptr));
// updating the parameters
// for the next thread
spos = epos + 1;
epos = spos + (N / T) - 1;
if (i == T - 2) {
epos = N - 1;
}
}
// Waiting for threads to finish
for (int i = 0; i < T; i++) {
int rc
= pthread_join(threads[i], NULL);
}
// Traverse the map
for (auto it: mp) {
// If frequency of current
// element is 1
if (it.second == 1) {
// Print the element
cout << it.first << " ";
}
}
}
// Drivers Code
int main()
{
// initializing the mutex lock
pthread_mutex_init(&lock1, NULL);
int T = 3;
int arr[] = { 1, 0, 5, 5, 2, 6 };
int N = sizeof(arr) / sizeof(arr[0]);
numbers_occuring_once(arr, N, T);
}
时间复杂度: O(N * log(N))
辅助空间: O(N)
注意:建议使用以下命令在基于Linux的系统中执行程序:
g++ -pthread program_name.cpp