循环中的线程池队列用户工作项 |拆分为子线程

Question

我有一个主线程，它从我的数据库收集活动作业以进行处理。 我想将此作业拆分为多线程，而不会出现任何崩溃或高 CPU 问题。

所以我选择了使用线程池的方式。我的主线程被另一个线程调用，工作原理如下：

    private void startChipDataManagment()
    {
        fillQueu();
        ThreadPool.SetMinThreads(_threadCount, _threadCount);
        ThreadPool.SetMaxThreads(_threadCount, _threadCount);
        do
        {
            for (int i = 0; i < MainJobsQueuList.Count; i++)
            {
                job CurrentJob = MainJobsQueuList[i];
                if (isSystemAlive())
                {
                    JobDataGenerationThread singleJobThread = new JobDataGenerationThread(CurrentJob.id, this);
                    ThreadPool.QueueUserWorkItem(new WaitCallback(singleJobThread.processJob), singleJobThread);
                    Thread.Sleep(0);
                }
            }

            while (_jobQueueCount > 0)
            {
                Thread.Sleep(10);
            }
            if (MainJobsQueuList.Count > 0)
            {
                Thread.Sleep(1);
                MainJobsQueuList.Clear();
                _jobQueueCount = 0;
            }
            Thread.Sleep(FILE_PROCESSED_AFTER_WAIT_TIME);
            fillQueu();
            
        } while (true);
     }

    private object syncLock = new object();
    public object syncLockChild = new object();
    int _jobQueueCount = 0;

    public void increaseThreadCount()
    {
        lock (syncLock)
        {
            _jobQueueCount++;
        }
        Console.WriteLine("Count increased. Current Count : {0}", _jobQueueCount);
    }
    public void decreaseThreadCount()
    {
        lock (syncLock)
        {
            _jobQueueCount--;
        }
        Console.WriteLine("Count decreased. Current Count : {0}", _jobQueueCount);
    }

我的 SingleJobThread 类如下：

    public JobDataGenerationThread(long assignedJobId, DataGenerationManagerThreadPoolNew mainThreadPool)
    {
        dbContextThread = new emvEntities();
        //currentJob = assignedJob;
        currentJobId = assignedJobId;
        CurrentAppSettings = new Settings();
        CurrentAppSettings.ReadSettings();

        SettingsManager SM = new SettingsManager();
        Deskey = SM.GetDesKey();

        _mainThreadPool = mainThreadPool;
    }

    public void processJob(object obj)
    {
        TimeSpan stTime = DateTime.Now.TimeOfDay;
        job SingleJob = /*currentJob;*/ GetSelectedJob(currentJobId);
        try
        {                
            _mainThreadPool.increaseThreadCount();
            /*Do some work.. Change flags of some Jobs*/
        }
        catch (Exception ext) {
            LogGenerator.WriteLog(NLog.LogLevel.Error, 1, LogGenerator.ModuleType.DataEngine, "I see Error : " + ext.ToString() + "MY JOB ID :" + currentJobId, true);
        }
        finally
        {
            try
            {
                SingleJob.isDataProcessingStarted = true;
                SingleJob.creationdate = DateTime.Now;
                dbContextThread.SaveChanges();
                _mainThreadPool.decreaseThreadCount();
                TimeSpan etTime = DateTime.Now.TimeOfDay;
                LogGenerator.WriteLog(NLog.LogLevel.Trace, 1, LogGenerator.ModuleType.DataEngine, string.Format("I have Taken total: {0} seconds. My JOB ID is : {1}", etTime - stTime, currentJobId), true);
                //Console.WriteLine("I have taken total : {0} seconds", etTime - stTime);
            }
            catch (Exception excpt)
            {
                LogGenerator.WriteLog(NLog.LogLevel.Error, 1, LogGenerator.ModuleType.DataEngine, string.Format("!!!!!!!!!!!!!!!!!FATAL ERROR!!!!!!!!!!!!!!!!! \r\n{0}", excpt.ToString()), true, excpt);
                Console.WriteLine("!!!!!!!!!!!!!!!!!FATAL ERROR!!!!!!!!!!!!!!!!! {0}\r\n", excpt.ToString());
            }
        }

但是我实现了这个逻辑，我不相信我为 ThreadPool 设置的 ThreadCount 是正确的，而且我也有一些问题，比如在循环中创建 JobDataGenerationThread 对象并在 ThreadPool 中调用它的函数是否可以？我的项目绩效没有增加。我错过了什么？

我将不胜感激任何输入.. 谢谢

编辑：

感谢 David-Browne 的回答，我更新了如下代码并删除了垃圾代码。所以这似乎比我之前的代码更安全。

            Parallel.ForEach(MainJobsQueuList, opt, job =>
            {
                if (isSystemAlive())
                {
                    JobDataGenerationThread singleJobThread = new JobDataGenerationThread(job.id, this);
                    singleJobThread.processJob();
                }
            });

我删除了获取对象的 processJob 函数。 然而，我付出了所有努力，我发现性能问题是因为我调用的另一个应用程序在支持多线程进程方面存在一些限制..

Answer 1

并行循环就像这样：

var opt = new ParallelOptions() { MaxDegreeOfParallelism = 4 };
Parallel.ForEach(MainJobsQueuList, opt, job => ProcessJob(job));

几乎替换所有代码。