基于任务组的线程池（SequentialThreadPool）_业界新闻

背景

考虑有多个任务，比如有5个数据采集设备，需要对每个设备采集的数据进行处理，单个设备的数据需要按照数据产生的顺序进行处理，每个设备产生的频率不一致。应该如何处理上述需求？

方案一：
多线程处理：对每个设备采用一个线程进行处理，确保了设备的数据顺序。
弊端：当设备增加时，浪费了大量线程；对于频率低的设备线程利用率低，对于频率高的设备处理能力可能不足。

方案二：
线程池处理：虽然能很好的利用资源，但无法保证单个设备的数据有序性。

方案三：
协程：协程能较好的处理此问题，但需要C++20支持，且无栈协程具有传染性。

基于上述情况，提出了一种按照任务组进行顺序执行的线程池。

简介

基于C++11实现的一个head-only的线程池。其区别于普通线程池在于：可以对任务进行分组，同一个组内的任务按照入队顺序进行执行，不同组之间的任务抢占式执行。线程资源平衡利用，避免出现某个线程利用率高，某个线程利用率低的问题。

使用

#include "SequentialThreadPool.hpp" #include <iostream> #include <sstream> int main() { const auto start = std::chrono::high_resolution_clock::now(); SequentialThreadPool pool(4); for (int task = 0; task < 30; ++task) { const int group = task % 5; // Task Grouping pool.enqueue(group, [=]() { const auto end = std::chrono::high_resolution_clock::now(); const auto duration = std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count(); const int sleep = task == 0 ? 200 : rand() % 100 + 50; std::stringstream ss; ss << "Time:" << duration << ", group:" << group << ", task:" << task << ", tid:" << std::this_thread::get_id() << ", sleep:" << sleep << "ms\n"; std::cout << ss.str(); std::this_thread::sleep_for(std::chrono::milliseconds(sleep)); }); } return 0; }

输出

Time:5, group:0, task:0, tid:23788, sleep:200ms Time:5, group:1, task:1, tid:26944, sleep:91ms Time:5, group:2, task:2, tid:27324, sleep:91ms Time:5, group:3, task:3, tid:21836, sleep:91ms Time:105, group:4, task:4, tid:27324, sleep:117ms Time:105, group:3, task:8, tid:26944, sleep:117ms Time:105, group:2, task:7, tid:21836, sleep:117ms Time:216, group:1, task:6, tid:23788, sleep:91ms Time:231, group:0, task:5, tid:27324, sleep:84ms Time:231, group:3, task:13, tid:26944, sleep:84ms Time:231, group:2, task:12, tid:21836, sleep:84ms Time:324, group:4, task:9, tid:21836, sleep:50ms Time:324, group:2, task:17, tid:23788, sleep:117ms Time:324, group:1, task:11, tid:27324, sleep:50ms Time:324, group:0, task:10, tid:26944, sleep:50ms Time:386, group:0, task:15, tid:27324, sleep:119ms Time:386, group:4, task:14, tid:21836, sleep:119ms Time:386, group:3, task:18, tid:26944, sleep:119ms Time:449, group:1, task:16, tid:23788, sleep:84ms Time:513, group:0, task:20, tid:26944, sleep:74ms Time:513, group:3, task:23, tid:21836, sleep:74ms Time:513, group:2, task:22, tid:27324, sleep:74ms Time:543, group:4, task:19, tid:23788, sleep:50ms Time:589, group:1, task:21, tid:21836, sleep:128ms Time:589, group:3, task:28, tid:27324, sleep:128ms Time:589, group:2, task:27, tid:26944, sleep:128ms Time:605, group:0, task:25, tid:23788, sleep:119ms Time:728, group:1, task:26, tid:23788, sleep:74ms Time:728, group:4, task:24, tid:21836, sleep:108ms Time:838, group:4, task:29, tid:21836, sleep:112ms