Discussion of article "MQL5 vs QLUA - Why trading operations in MQL5 are up to 28 times faster?" - page 6

 
Renat Fatkhullin:

Only MT5 has nothing to do with these delays.

It was the exchange that took so long to respond.

That's strange, that's what the exchange replied:

Добрый вечер!

Задержки на нашей стороне по данным ордерам были маленькими.
(Latency пишется в микросекундах)

Date      Time           Login Latency  id_ord
Sep 14  17:30:15.0227000 ххххх 410 22891096373
Sep 14  17:30:12.6450710 ххххх 439 22891080332
Sep 14  17:30:16.2887880 ххххх 432 22891106171
Sep 14  17:30:16.3533160 ххххх 616 22891106450
Sep 14  17:30:05.7216320 ххххх 462 22891025944

С уважением,
Антон Водопьянов
Техническая поддержка ПАО Московская Биржа
help@moex.com

 
 
prostotrader:

That's odd, that's what the exchange replied:

Thank you, we'll request details of the server logs ourselves tomorrow.

Was it in Open? Please specify the correspondence between orders in MT5 and submitted exchange tickets.

 
Renat Fatkhullin:

Thanks, we'll ask for details of the server logs tomorrow.

This was in Discovery?

Yes, Discovery.

Added.

Accsess server IV

87.118.239.81:443

Accsess server IV 87.118.1189.81:443

Added

BR-12.16 ордер в MT5 - 44309414 (биржевой номер 22891025944 )
BR-11.16 ордер в MT5 - 44309578 (биржевой номер 22891080332 )
BR-11.16 ордер в MT5 - 44309644 (биржевой номер 22891096373 )
BR-12.16 ордер в MT5 - 44309680 (биржевой номер 22891106171 )
BR-11.16 ордер в MT5 - 44309682 (биржевой номер 22891106450 )
 
Renat Fatkhullin:

Thanks, we'll ask for details of the server logs tomorrow.

Was it in Open? Please specify the correspondence between orders in MT5 and submitted stock tickets.


Sorry, didn't wait for the result.

Can we find out who is lagging?

 
Sergey Chalyshev:

Sorry, I didn't get the results.

Can we find out who's lagging?


https://www.mql5.com/ru/forum/38456/page80#comment_2835361

ФОРТС. Вопросы по исполнению
ФОРТС. Вопросы по исполнению
  • www.mql5.com
С большими проблемами удалось это сделать (начальник отдела по работе с профессиональными клиентами ДЦ Открытие Евгений Сергеевич,.
 

Are QLUA and Lua the same thing?

(no subject)
(no subject)
  • 2014.07.20
  • man_of_motley
  • man-of-motley.livejournal.com
История (к сожалению, больше профессиональная) к тому что большие корпорации не так уж мега-разумны как кажется некоторым и эффективность их работы не так уж высока. На примере Google с их V8 и вообще на тему интеграции скриптовых языков. Для начала немножко вводной: Так уж получилось что одно из наших флагманских приложений требует дать...
 

Yes.

The text on your link is misleading and creates an illusion that Lua is normal and fast. In fact it is sufficient as a primitive scripting of system calls with complete abandonment of its own maths and heavy calculations.

In computer games, it is used as a cheap way to implement flexible scripting of scenes that can be taken out of the project code and given to external designers/scriptors. On this basis, these developers suddenly start saying that Lua is a normal and fast language. Of course, it is not.

If we talk about maths and calculations, it is 50 to 600 times faster than MQL5.
 
Renat Fatkhullin:

Yes.

The text on your link is misleading and creates an illusion that Lua is normal and fast. In fact it is sufficient as a primitive scripting of system calls with complete abandonment of its own maths and heavy calculations.

In computer games, it is used as a cheap way to implement flexible scripting of scenes that can be taken out of the project code and given to external designers/scriptors. On this basis, these developers suddenly start saying that Lua is a normal and fast language. Of course, it is not.

If we talk about maths and calculations, it is 50 to 600 times faster than MQL5.

I read diagonally and looked at the source text in Lua.


It is clear that the text is advertising. It is difficult to say how the speed of order placement relates to the speed of the lua machine. The state of the infrastructure plays a role here, not microseconds of difference in execution time on the local processor.

Regarding the lua text. Renate. Lua does not imply the programming style familiar to C++ or other delphi. Before writing tests in a programming language and raising yourself to the rank of a guru, you should have some experience of programming in this language. You don't have it.


Concerning performance.

Since the moment of embedding lua 5.1 support in the terminal, quik offers 2 options for creating embedded scripts.

1. Simple and undemanding to hardware resources. In this case these scripts are written in lua, which CARDINALLY saves programming costs (stomping buttons) compared to any C-like language. In this case we save development time, and we are not interested in performance and saving milliseconds at runtime.

2. resource-intensive and with low-level access. In this case we write in pure C++, executing the algorithm as a lua extension module. In this case we get our hands on all the features of VC++ or something similar. I don't think you have the guts to claim that mql5, all other things being equal, can be faster or provide more convenience or features ).


In short, everything written in the original article is bubble blowing. You are developers of some sort, not forex firms for noodling. Or do I think too well of your firm?

 
s_mike:

I read diagonally and looked at the source text in Lua.


It is clear that the text is advertising. It is hard to say how the speed of bidding relates to the speed of the lua machine. The state of the infrastructure plays a role here, not microseconds of difference in execution time on the local processor.

Regarding the lua text. Renate. Lua does not imply the programming style familiar to C++ or other delphi. Before writing tests in a programming language and raising yourself to the rank of a guru, you should have some experience of programming in this language. You don't have it.

Read the text carefully, not diagonally.

I'm afraid I'm sure that you have no programming experience to challenge both the tests themselves and their results.


Concerning performance.

quik since the moment of embedding lua 5.1 support in the terminal offers 2 options for creating embedded scripts.

1. Simple and undemanding to hardware resources. In this case these scripts are written in lua, which CARDINALLY saves programming costs (stomping buttons) compared to any C-like language. In this case we save development time, and we are not interested in performance and saving milliseconds at runtime.

It doesn't save anything. It is a miserable, catastrophically limited and slow to the point of madness language.

It is generally unsuitable for financial and data processing. Its task is to be a simple script binding if/then conditions between other high-level calls. That's how it was created.


2. resource-intensive and with low-level access. In this case we write in pure C++, executing the algorithm as a lua extension module. In this case we get our hands on all the features of VC++ or something similar. I think you don't have the guts to claim that mql5, all other things being equal, can be faster or provide more convenience or features ).

And inheriting all the brakes and limitations of Lua. Here is not to extract and transfer tonnes of source data to lua, then to the library, and then back and then to the terminal. There the losses will be thousand-fold compared to MQL5, where all the data is directly available to MQL5. MQL5 is deeply integrated into the terminal kernel, all the processes of which are built to serve MQL5 programmes.

Besides, we have already shown that MQL5 works at the C++ level.

In short, everything written in the original article is just blowing bubbles. You are developers, not forex firms for noodling. Or do I think too highly of your firm?

You're just a theorist with no practical knowledge.

Not understanding the concept of velocity puts you out of the trading game:

  • 5 times faster to update the stack? Doesn't matter!
  • 28 times faster to execute trades? So what?
  • MQL5 is600 times faster than QLUA? Come on, it's just blowing bubbles!

It's funny to listen to couch warriors who try to argue with experts on their fingers. When we release these kinds of tests, we go through our own criticism first, knowing perfectly well the consequences of publishing weak evidence.

In all the time, not a single person has been able to prove or show the errors in this paper. Proofs should be with reproducible code and survive criticism from specialists.

Битва за скорость: QLUA vs MQL5 - почему MQL5 быстрее от 50 до 600 раз?
Битва за скорость: QLUA vs MQL5 - почему MQL5 быстрее от 50 до 600 раз?
  • smart-lab.ru
Для сравнения языков MQL5 и QLUA мы написали несколько тестов, которые замеряют скорость выполнения базовых операций. В тестах использовался компьютер с Windows 7 Professional 64 bit, MetaTrader 5 build 1340 и QUIK версии 7.2.0.45. Результаты представлены в таблице, где все значения представлены в миллисекундах (чем меньше время, тем лучше...
 
Renate.

Why are you so excited? Hit a sore spot? You'll have to forgive me, couch warrior)))

Regarding my person:
I have written enough letters in files with the extension .lua and .mq5. I have written more letters in lua than you have. By orders of magnitude.

And if I claim that your style of programming in lua is amateurish, it means that there are grounds for it.

Now to the point.


I repeat. If you need the speed of calculations, you can simply write a module in the same Visual Studio, to which all the necessary raw data is passed and the processing goes inside this module). Google will help you find examples of such ready modules, including very popular ones.

What huge arrays do you need to bullet from lua into a C++ module? Stop fantasising. A unit of data is a table of a dozen short fields. It costs 0 milliseconds to transfer them to the module through the lua api pseudo-stack.

Any nonsense about taking me out of the trading boundaries and other if/else wrappings should be referred to a psychiatrist.


Mikhail. www.<Removed by Artyom Trishkin>. ru