c# - MaxDegreeOfParallelism 为 1 的 for 循环和 Parallel.For() 之间的性能损失

Question

我想做如下的事情:

int firstLoopMaxThreads = 1; // or -1
int secondLoopMaxThreads = firstLoopMaxThreads == 1 ? -1 : 1;

Parallel.For(0, m, new ParallelOptions() { MaxDegreeOfParallelism = firstLoopMaxThreads }, i =>
{
    //do some processor and/or memory-intensive stuff
    Parallel.For(0, n, new ParallelOptions() { MaxDegreeOfParallelism = secondLoopMaxThreads }, j =>
    {
        //do some other processor and/or memory-intensive stuff
    });
});

当 secondLoopMaxThreads = 1 时，将内部 Parallel.For 循环与普通 for 循环交换是否值得，从性能角度来看？常规 for 循环与 Parallel.For 循环(MaxDegreeofParallelism = 1)之间的性能差异是什么？

Answer 1

这取决于您谈论的迭代次数以及您谈论的性能水平，以回答是否值得。在您的上下文中，1 毫秒被认为是很多还是很少？

我做了一个基本的测试如下(因为 Thread.Sleep 并不完全准确.. 尽管 for 循环每次测量 15,000ms 到 1ms 以内)。超过 15,000 次迭代重复 5 次，与标准 for 循环相比，它通常会增加大约 4 毫秒的开销……但当然结果会因环境而异。

for (int z = 0; z < 5; z++)
{
    int iterations = 15000;

    Stopwatch s = Stopwatch.StartNew();

    for (int i = 0; i < iterations; i++)        
        Thread.Sleep(1);        

    s.Stop();

    Console.WriteLine("#{0}:Elapsed (for): {1:#,0}ms", z, ((double)s.ElapsedTicks / (double)Stopwatch.Frequency) * 1000);

    var options = new ParallelOptions() { MaxDegreeOfParallelism = 1 };
    s = Stopwatch.StartNew();

    Parallel.For(0, iterations, options, (i) => Thread.Sleep(1));

    s.Stop();

    Console.WriteLine("#{0}: Elapsed (parallel): {1:#,0}ms", z, ((double)s.ElapsedTicks / (double)Stopwatch.Frequency) * 1000);
}