A little program help me to know about Parallel Class.
Task is a good ,amazing thing. Background : There are two serial methods is the excessive cost of time in my work . Leader hope I can make it sense, do some optimization and runtime performance. So I study Task in comparison with Parallel.
Result ：I chose Task as a solution. Compare time, both almost. But I think Parallel is more useful in a special environment, such as a 10000 loop. Normally,one thread loops 10000 times, C# call cpu cores threads to do the thing at same time if in Parellel. (并行用在一些特定的环境，比如一个10000次的循环，正常的情况是用1个线程完成这10000次，而如果用并行循环， c#会用cpu内核数相等的线程来共同完成这10000次。) There is an interesting thing in Parallel. When to use a GLOBAL varable or a LOCAL varable ,get very different results.
CODE:
namespace ParallelTasks { using System; using System.IO; using System.Linq; using System.Text; using System.Threading; using System.Threading.Tasks; using System.Net; using System.Diagnostics;
class ParallelInvoke
{
static void Main()
{
//1.LearnTask
//var taskTop = new Task(() =>
//{
// Thread.Sleep(500);
// Console.WriteLine("任务taskTop运行在线程“{0}”上",
// Thread.CurrentThread.ManagedThreadId);
//});
//var taskMiddle = new Task(() =>
//{
// Thread.Sleep(500);
// Console.WriteLine("任务taskMiddle运行在线程“{0}”上",
// Thread.CurrentThread.ManagedThreadId);
//});
//var taskBottom = new Task(() =>
//{
// Thread.Sleep(500);
// Console.WriteLine("任务taskBottom运行在线程“{0}”上",
// Thread.CurrentThread.ManagedThreadId);
//});
//taskTop.Start();
//taskMiddle.Start();
//taskBottom.Start();
//Task.WaitAll(new Task[] { taskTop, taskMiddle, taskBottom });
//2. Compare Parllel to Task
//ParallelDemo p = new ParallelDemo();
//p.ParallelDemoRun();
//3. Compare Parllel with For
ParallelMethod p = new ParallelMethod();
p.Method();
Console.ReadKey();
}
public class ParallelDemo
{
private Stopwatch stop = new Stopwatch();
public void Run1()
{
Thread.Sleep(2000);
Console.WriteLine("Task1 run 2s");
}
public void Run2()
{
Thread.Sleep(3000);
Console.WriteLine("Task2 run 3s");
}
public void ParallelDemoRun()
{
stop.Start();
Parallel.Invoke(Run1, Run2);
stop.Stop();
Console.WriteLine("Parallel uses " + stop.ElapsedMilliseconds + "ms"); // about 3013ms
stop.Restart();
Run1();
Run2();
stop.Stop();
Console.WriteLine("Normal run " + stop.ElapsedMilliseconds + "ms"); //=4999ms
var task1 = new Task(() => Run1());
var task2 = new Task(() => Run2());
stop.Restart();
task1.Start();
task2.Start();
Task.WaitAll(new Task[] { task1, task2 });
stop.Stop();
Console.WriteLine("WaitAll run " + stop.ElapsedMilliseconds + "ms");//=2999s
}
}
public class ParallelMethod
{
private Stopwatch stop = new Stopwatch();
public void Method()
{
//** global :Global Varibale **
int global =0;
stop.Restart();
Parallel.For(0, 10000, item =>
{
//** sum : Local Variable **
//int sum = 0;
for (int j = 0; j < 60000; j++)
{
global++;
// sum++;
}
});
stop.Stop();
//** Local Variable takes 328ms **
//** Global Variable takes 3002ms **
Console.WriteLine("Parallel run " + stop.ElapsedMilliseconds + "ms");
stop.Restart();
for (int i = 0; i < 10000; i++)
{
//int sum = 0;
for (int j = 0; j < 60000; j++)
{
global++;
// sum++;
}
}
stop.Stop();
//** Local Variable takes 1168ms**
//** Global Variable takes 1505ms **
Console.WriteLine("Normal run " + stop.ElapsedMilliseconds + "ms");
}
}
}
}