MT5和速度在行动

fxsaber 2020.10.08 06:04 #491

fxsaber, 2020.10.07 11:13

#include <fxsaber\Benchmark\Benchmark.mqh> // https://www.mql5.com/ru/code/31279

void OnTick()
{
  _BV(Print(""), 1);
}

无法检查它在慢速VPS上的表现。

如果你在OrderSend之后做任何打印输出（例如订单发送结果），最好将它们累积在一个字符串变量中，并在退出On-function时将其发送给打印。

否则，你可能会失去很多连接到打印-快照的钱。多少钱 - 你可以回答，如果有人在他的VPS上运行上述顾问。

ZZZ 最简单的实现。

#include <fxsaber\Benchmark\Benchmark.mqh> // https://www.mql5.com/ru/code/31279

class PRINT
{
private:
  string Str;
  
public:
  PRINT() : Str(NULL) { ::StringSetLength(this.Str, 5000); }
  ~PRINT() { if (this.Str != NULL) ::Print(this.Str); }
  
  template <typename T>
  void Add( const T PrintStr ) { this.Str += ((this.Str == NULL) ? NULL : "\n") + (string)PrintStr; }
};

#define Print PrintObj.Add // Закомментируйте, чтобы посмотреть скорость обычного Print.

void OnTick()
{  
  PRINT PrintObj; // Накопитель Print-ов.
  
  for (int i = 0; i < 5; i++)
  {
    _BV(Print(i), 1)
    
    Sleep(100); // Эмуляция OrderSend
  }
}

在快速的机器上，这种解决方案要比通常的方法快得多。在VPS上怎么做--不知道。

Edgar Akhmadeev 2020.10.08 06:53 #492

fxsaber:

如果你在OrderSend之后做任何打印输出（例如订单发送结果），最好将它们累积在一个字符串变量中，并在退出On-function时将其发送给打印。

否则，你可能会失去很多连接到打印-快照的钱。多少钱 - 你可以回答，如果有人在他的VPS上运行上述顾问。

在我的3美元VPS上

2020.10.08 09:50:59.631 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 3121 mсs.
2020.10.08 09:50:59.724 Test Print (EURUSD,H1)  
2020.10.08 09:50:59.724 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 78 mсs.
2020.10.08 09:50:59.837 Test Print (EURUSD,H1)  
2020.10.08 09:50:59.837 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 86 mсs.
2020.10.08 09:50:59.930 Test Print (EURUSD,H1)  
2020.10.08 09:50:59.931 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 159 mсs.
2020.10.08 09:51:00.036 Test Print (EURUSD,H1)  
2020.10.08 09:51:00.036 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 84 mсs.
2020.10.08 09:51:00.138 Test Print (EURUSD,H1)  
2020.10.08 09:51:00.138 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 94 mсs.
2020.10.08 09:51:00.243 Test Print (EURUSD,H1)  
2020.10.08 09:51:00.243 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 86 mсs.
2020.10.08 09:51:00.380 Test Print (EURUSD,H1)  
2020.10.08 09:51:00.380 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 54 mсs.
2020.10.08 09:51:00.973 Test Print (EURUSD,H1)  
2020.10.08 09:51:00.973 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 79 mсs.
2020.10.08 09:51:01.129 Test Print (EURUSD,H1)  
2020.10.08 09:51:01.129 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 91 mсs.
2020.10.08 09:51:01.186 Test Print (EURUSD,H1)  
2020.10.08 09:51:01.186 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 60 mсs.
2020.10.08 09:51:01.320 Test Print (EURUSD,H1)  
2020.10.08 09:51:01.321 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 75 mсs.
2020.10.08 09:51:01.402 Test Print (EURUSD,H1)  
2020.10.08 09:51:01.402 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 86 mсs.
2020.10.08 09:51:01.546 Test Print (EURUSD,H1)  
2020.10.08 09:51:01.555 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 72 mсs.
2020.10.08 09:51:02.158 Test Print (EURUSD,H1)  
2020.10.08 09:51:02.158 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 67 mсs.
2020.10.08 09:51:04.651 Test Print (EURUSD,H1)  
2020.10.08 09:51:04.651 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 74 mсs.
2020.10.08 09:51:04.929 Test Print (EURUSD,H1)  
2020.10.08 09:51:04.930 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 200 mсs.
2020.10.08 09:51:07.200 Test Print (EURUSD,H1)  
2020.10.08 09:51:07.200 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 58 mсs.
2020.10.08 09:51:07.330 Test Print (EURUSD,H1)  
2020.10.08 09:51:07.330 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 68 mсs.
2020.10.08 09:51:07.452 Test Print (EURUSD,H1)  
2020.10.08 09:51:07.460 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 77 mсs.
2020.10.08 09:51:07.728 Test Print (EURUSD,H1)  
2020.10.08 09:51:07.728 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 64 mсs.
2020.10.08 09:51:07.870 Test Print (EURUSD,H1)  
2020.10.08 09:51:07.870 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 72 mсs.
2020.10.08 09:51:07.967 Test Print (EURUSD,H1)  
2020.10.08 09:51:07.967 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 58 mсs.
2020.10.08 09:51:08.130 Test Print (EURUSD,H1)  
2020.10.08 09:51:08.130 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 56 mсs.
2020.10.08 09:51:11.274 Test Print (EURUSD,H1)  
2020.10.08 09:51:11.275 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 63 mсs.
2020.10.08 09:51:11.409 Test Print (EURUSD,H1)  
2020.10.08 09:51:11.409 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 73 mсs.
2020.10.08 09:51:11.903 Test Print (EURUSD,H1)  
2020.10.08 09:51:11.903 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 86 mсs.
2020.10.08 09:51:12.022 Test Print (EURUSD,H1)  
2020.10.08 09:51:12.022 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 79 mсs.
2020.10.08 09:51:12.699 Test Print (EURUSD,H1)  
2020.10.08 09:51:12.700 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 69 mсs.
2020.10.08 09:51:12.977 Test Print (EURUSD,H1)  
2020.10.08 09:51:12.977 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 67 mсs.
2020.10.08 09:51:13.226 Test Print (EURUSD,H1)  
2020.10.08 09:51:13.226 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 98 mсs.
2020.10.08 09:51:13.412 Test Print (EURUSD,H1)  
2020.10.08 09:51:13.412 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 93 mсs.
2020.10.08 09:51:13.854 Test Print (EURUSD,H1)  
2020.10.08 09:51:13.854 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 59 mсs.
2020.10.08 09:51:14.000 Test Print (EURUSD,H1)  
2020.10.08 09:51:14.000 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 62 mсs.
2020.10.08 09:51:18.343 Test Print (EURUSD,H1)  
2020.10.08 09:51:18.343 Test Print (EURUSD,H1)  Alert: Time[Test Print.mq5 5: Print()] = 86 mсs.

fxsaber 2020.10.08 06:55 #493

Edgar Akhmadeev:

在我的3美元VPS上。

谢谢你。绝对是一个更好的变通办法。

Anton 2020.10.08 07:48 #494

fxsaber:

#include <fxsaber\Benchmark\Benchmark.mqh> // https://www.mql5.com/ru/code/31279

long GetAnotherChart()
{
  long Chart = ::ChartFirst();
  
  while (Chart == ChartID())
    Chart = ChartNext(Chart);
 
  return(Chart);     
}


void OnTick()
{  
  const long Chart = GetAnotherChart();
  
  if (Chart)
    _B(EventChartCustom(Chart, 123, 0, 0, NULL), 1);
  
  _B(EventChartCustom(0, 123, 0, 0, NULL), 1);
}

寄给别人的图表比寄给自己的更贵。

这样试试吧。

_B(EventChartCustom(-1, 123, 0, 0, NULL), 1);

fxsaber 2020.10.08 08:21 #495

Anton:

这样试试吧。

// https://www.mql5.com/ru/forum/342090/page50#comment_18647171
class PRINT
{
private:
  string Str;
public:
  PRINT() : Str(NULL) { ::StringSetLength(this.Str, 5000); }
  ~PRINT() { if (this.Str != NULL) ::Print(this.Str); }
  
  template <typename T>
  void Add( const T PrintStr ) { this.Str += ((this.Str == NULL) ? NULL : "\n") + (string)PrintStr; }
};

#define Alert PrintObj.Add

PRINT PrintObj;

//#define BENCHMARK_OFF // Выключение замеров.
#include <fxsaber\Benchmark\Benchmark.mqh> // https://www.mql5.com/ru/code/31279

void OnTick()
{  
  _B(EventChartCustom(-1, 123, GetMicrosecondCount(), 0, NULL), 1);
}

void OnChartEvent(const int id, const long& lparam,const double& dparam,const string& sparam)
{
  if (id == 1123)
    Print("Time[from OnTick To OnChartEvent] = " + (string)(GetMicrosecondCount() - lparam) + " mcs.");
}

结果。

2020.10.08 11:35:52.050 Benchmark.mqh: TimeAvg[GetMicrosecondsCount()] = 0 mcs.
2020.10.08 11:35:52.050 Benchmark.mqh is On.
2020.10.08 11:35:55.617 Time[from OnTick To OnChartEvent] = 14 mcs.
2020.10.08 11:35:55.821 Time[from OnTick To OnChartEvent] = 3 mcs.
2020.10.08 11:35:56.729 Time[from OnTick To OnChartEvent] = 4 mcs.
2020.10.08 11:35:56.932 Time[from OnTick To OnChartEvent] = 9 mcs.
2020.10.08 11:35:57.841 Time[from OnTick To OnChartEvent] = 9 mcs.
2020.10.08 11:35:58.177 Bench_Stack = 0, 1 <= Time[Test9.mq5 23 in OnTick: EventChartCustom(-1,123,GetMicrosecondCount(),0,NULL)] = 2 mcs.
2020.10.08 11:35:58.177 Bench_Stack = 0, 1 <= Time[Test9.mq5 23 in OnTick: EventChartCustom(-1,123,GetMicrosecondCount(),0,NULL)] = 2 mcs.
2020.10.08 11:35:58.177 Bench_Stack = 0, 1 <= Time[Test9.mq5 23 in OnTick: EventChartCustom(-1,123,GetMicrosecondCount(),0,NULL)] = 2 mcs.

当测光被关闭时。

2020.10.08 11:15:19.084 Benchmark.mqh: TimeAvg[GetMicrosecondsCount()] = 0 mcs.
2020.10.08 11:15:19.084 Benchmark.mqh is Off.
2020.10.08 11:15:24.171 Time[from OnTick To OnChartEvent] = 4 mcs.
2020.10.08 11:15:24.376 Time[from OnTick To OnChartEvent] = 3 mcs.
2020.10.08 11:15:27.010 Time[from OnTick To OnChartEvent] = 3 mcs.
2020.10.08 11:15:31.859 Time[from OnTick To OnChartEvent] = 5 mcs.
2020.10.08 11:15:32.266 Time[from OnTick To OnChartEvent] = 4 mcs.
2020.10.08 11:15:32.470 Time[from OnTick To OnChartEvent] = 3 mcs.
2020.10.08 11:15:32.670 Time[from OnTick To OnChartEvent] = 6 mcs.

也就是说，切换到 "OnEmpty "需要不到10µs。

为什么不对零做同样的事？

Anton 2020.10.08 09:00 #496

fxsaber:

也就是说，切换到 "OnEmpty "需要不到10µs。

为什么不对零做同样的事？

不同的意义和不同的执行机制。0 - 事件为 "自己的 "图表。-1 - 事件在自己的图表队列中。

fxsaber 2020.10.08 14:59 #497

关于交易、自动交易系统和交易策略测试的论坛

MT5和速度在行动

fxsaber, 2020.10.07 12:41

尽管我一再要求，但我从未尝试在MQL5中使用TimeCurrentMsc。

会是这样吗？

fxsaber 2020.10.08 15:27 #498

Anton:

不同的意义和不同的执行机制。0是 "自己 "图表队列中的一个事件。-1 - 事件在自己的队列中。

事实证明，通过-1的异步不能实现相同的警报。

A100 2020.10.08 20:57 #499

fxsaber:

这意味着切换到OnEmpty需要不到10微秒的时间。

正如你所看到的，在你的帮助下，一切都解决得比最乐观的预测要快，原来有一个隐藏的后门。

Roman 2020.10.08 21:22 #500

fxsaber:

会有吗？

不太一样，但由mcs间隔的计数器。
是的，TimeCurrentMcs 会更好，我加入了这个愿望。

//+------------------------------------------------------------------+
//| Script program start function                                    |
//+------------------------------------------------------------------+
void OnStart()
{
   ulong mcs = 0;
   
   while(!_StopFlag)
   {
      mcs = GetMicrosecondCount();
     
      switch(GetInterval(mcs, 5000000)) //5 секунд
      { 
         case 1: 
            Print((string) mcs); 
            break;
         case 0:       
            break;             
      }    
   }
}

//--------------------------------------------------------------------
//Получить интервал 
ulong prevCount = 0;
int GetInterval(ulong currCount, ulong mcsIntrval)
{ 
   int res = 0;   
   switch((currCount - prevCount) > mcsIntrval) 
   { 
      case 1: 
         prevCount = currCount;
         res = 1; 
         break; 
      default:
         break;   
   } 
   
   return(res);
}

2020.10.09 00:15:45.712 TestScript (MNQZ20,M1)  5000001
2020.10.09 00:15:50.712 TestScript (MNQZ20,M1)  10000002
2020.10.09 00:15:55.712 TestScript (MNQZ20,M1)  15000003
2020.10.09 00:16:00.712 TestScript (MNQZ20,M1)  20000004
2020.10.09 00:16:05.712 TestScript (MNQZ20,M1)  25000006
2020.10.09 00:16:10.712 TestScript (MNQZ20,M1)  30000007
2020.10.09 00:16:15.712 TestScript (MNQZ20,M1)  35000008
2020.10.09 00:16:20.712 TestScript (MNQZ20,M1)  40000009
2020.10.09 00:16:25.713 TestScript (MNQZ20,M1)  45000581
2020.10.09 00:16:30.713 TestScript (MNQZ20,M1)  50000582

MT5和速度在行动 - 页 50