Parallel process interest in C#

Question

In web api controller: I am a newer in parallel and Async&Task programming. Need someone help me code review.

public async Task<IHttpActionResult> DoWork()
{
    List<string> listTask = await Helper.GetBlockBlock();
    userIds = listTask;
    this.ParallelTaskWork();//step 1,must be finished,but I an not sure when..
    await this.Service.GetAllClientInterestInfoList();//executing after step 1 
    return this.Ok();
}
public async void ParallelTaskWork()
{
    Task[] taskList = this.TaskMaker(totalNum);
}
public Task[] TaskMaker(int threadNum)
{
    Task[] taskList = new Task[threadNum];
    Parallel.For(0, taskList.Length, i => taskList[i] = new Task(this.ProcessDealTask));
    Parallel.ForEach(taskList, t => t.Start());
    return taskList;
}
private void ProcessDealTask()
{
    for (int i = 0; i < totalNum; i++)
    {
        List<string> users = this.GetPartitionOfList(userIds, i, totalNum);
        Task task = this.ParallelDealUserList(users);
        task.Wait();
    }
}
private List<T> GetPartitionOfList<T>(List<T> userList, int index, int total)
{
    int section = (userList.Count / total) + 1;
    int end = (index + 1) * section > userList.Count ? userList.Count - index * section : section;
    if (end < 0) end = 0;
        int start = index * section < userList.Count ? index * section : userList.Count - 1;
    if (start < 0) start = 0;
        return userList.GetRange(start, end);
}
private async Task ParallelDealUserList(IEnumerable<string> list)
{
    await Task.Run(() => Parallel.ForEach(list, async (id, loopState) =>
    {
        try
        {
            await GrainClient.GrainFactory.GetGrain<IUser>(id.ToGuid()).DoProcessInterestAsync(true);
        }
        catch (Exception e)
        {
        }
    }));
}

Any performance problem? The method ParallelDealUserList I refer to https://stackoverflow.com/questions/19102966/parallel-foreach-vs-task-run-and-task-whenall somebody said

there is a disadvantage to use Task.Run in a loop- With Parallel.ForEach, there is a Partitioner which gets created to avoid making more tasks than necessary

and The method TaskMaker https://stackoverflow.com/questions/22988630/parallel-for-w-action-that-creates-different-tasks-c-sharp

Answer 1

First thing is that you should limit the degree of parallelism, to ex: 4 threads

Parallel.ForEach(
    listOfWebpages,
    new ParallelOptions { MaxDegreeOfParallelism = 4 },
    webpage => { Download(webpage); }
);

Without this, you are going to DDoS your service or resource you are using.

---

You seem to be using Project Orleans and AFAIK a grain should be a single threaded container. You should be able to do just:

private void ParallelDealUserList(IEnumerable<string> list)
{
    foreach(var item in list)
        GrainClient
            .GrainFactory
            .GetGrain<IUser>(item.id.ToGuid())
            .DoProcessInterestAsync(true);
}

---

I am not all that familiar with Project Orleans. Most likely this can be used, but I would use microservices (aka SOA2) instead.

Answer 2

It is really dangerous to fork background threads in ASP.NET. Task Scheduler immediately gets crazy, which degrades performance significantly.

There is a way though: HostingEnvironment.QueueBackgroundWorkItem.

QBWI schedules a task which can run in the background, independent of any request. This differs from a normal ThreadPool work item in that ASP.NET automatically keeps track of how many work items registered through this API are currently running, and the ASP.NET runtime will try to delay AppDomain shutdown until these work items have finished executing.

Please have a look at How to run Background Tasks in ASP.NET