Questions from Beginners MQL5 MT5 MetaTrader 5 - page 68

 
Yedelkin:
In first case a variable of double type is returned immediately, in second case a variable of MqlTick structure type (five elements of different types) is returned.

I wanted to understand - which variant is recommended by Developers (are they completely identical in terms of speed and obtained values)

 
A100: I wanted to know which option the Developers recommend (are they completely identical in terms of speed and resulting values)
Ok, we'll wait for their response. I - so it's more convenient to use the first option, for mining one asc.
 
technotrance:
Hello!
I have some questions about the agents.

1. I have a total of 128 MetaTester 5 agents running. Some of them are running on very powerful machines with Intel Xeon E5645, Intel Xeon E5649, Intel Xeon E5-2620 processors...Why do these agents have such a low PR? They have about 100 (+ -). At the same time, I have weaker machines with IntelXeon E3-1240CPUs -- their PR=140. Why is it so? Because performance doesn't depend only on processor frequency.

The secret is that no Xeon is in any way faster than a desktop processor. On the contrary server processors are much weaker than desktop ones. And no megacaches are of any help to them.

Understanding this is discouraging.

2. If I believe information on payments, which is written here: https://cloud.mql5.com/ru/faq/payments , then my 128 agents must accumulate at least 10 credits per day. In reality, however, I have accumulated for the day at most 0.25 credits. Why is this so?

Look at agent load - most likely it's not even up to 1%, as there aren't enough tasks for everyone.
 
Renat:

The secret is that no Xeon is in any way faster than a desktop processor. On the contrary, server processors are much weaker than desktop ones. And no megacaches are of any help to them.

Understanding this is discouraging.

Look at agent load - most likely it's not even up to 1%, as there are not enough tasks for everyone.
Renat, implement the same "cloud" in mcl4, people get tired of optimizing for several days and you and us feel good...
 
Renat:
Look at the agent load - it's probably not even 1% because there aren't enough tasks for everyone.
Where can I look at this? And, for my part, is there anything I can do to make these very tasks go more to me and not to others? :)
 
technotrance:
Where can I see it? And for my part, is there anything I can do to make those same tasks go more to me than to others? :)

"Why, I want them to come to me, too."

There's no need to tug at the quilt with the agents. Everything is distributed fairly.

 
The demo account does not open. I wonder why.
Документация по MQL5: Стандартные константы, перечисления и структуры / Состояние окружения / Информация о счете
Документация по MQL5: Стандартные константы, перечисления и структуры / Состояние окружения / Информация о счете
  • www.mql5.com
Стандартные константы, перечисления и структуры / Состояние окружения / Информация о счете - Документация по MQL5
 
papaklass:
Increase the number of agents by an order of magnitude and raise their PR.
There are already 128 of them... And the most powerful servers are all already in use, but as it turned out, xeons don't count...
 
A100:

is there any fundamental difference

if only Ask is required, for example

If you only need Ask, there is no fundamental (or rather no) difference between

double ask=SymbolInfoDouble(_Symbol,SYMBOL_ASK);

и

double  ask=EMPTY_VALUE;
MqlTick tick;
//---
if(SymbolInfoTick(_Symbol,tick))
   ask=tick.ask;

no.

If both Ask and Bid are needed, then

double ask=SymbolInfoDouble(_Symbol,SYMBOL_ASK);
double bid=SymbolInfoDouble(_Symbol,SYMBOL_BID);

may give results different from

double  bid=EMPTY_VALUE;
double  ask=EMPTY_VALUE;
MqlTick tick;
//---
if(SymbolInfoTick(_Symbol,tick))
  {
   bid=tick.bid;
   ask=tick.ask;
  }

because in the first case no one guarantees that Ask and Bid belong to the same tick.

 
How to get in Expert Advisor the values of indicator that shows in future, e.g. red bar https://www.mql5.com/ru/code/129? CopyBuffer gives only on 0 candle, the rest is DBL_MAX.
Авторегрессивная модель (AR) экстраполяции цен
Авторегрессивная модель (AR) экстраполяции цен
  • votes: 15
  • 2010.07.05
  • gpwr
  • www.mql5.com
Этот индикатор использует авторегрессивную модель для экстраполяции будущих цен.