OpenCl and the tools for it. Reviews and impressions. - page 18
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Oh, pity there's no likes! Strongly put, I just didn't have the eloquence.
No offence to the developers. But really, I think there is something to think about. The marketing component is a bit overdone, IMHO.
Please don't get confused and don't go into universal "we are offended" mode when it comes to technical stuff.
Reshetov blatantly started to bullshit and got an adequate response with technical arguments.
Your case. Go to ninja (C#) or Dukas (J++) if you want comparisons so much. Look at the problems of ordinary users, the platforms' capabilities, their forums, tech-support.
Or do you want a platform with the simplest language (or even with visual code generation)? Yes, there are such. But I'm afraid they won't satisfy you in terms of their capabilities.
It's a common trend, and there's no getting away from it. You still have to sacrifice, since it is the riskiest business in the world.
And once again: in MQL5 you can write in an almost conventional procedural style, almost without OOP. If the code is not very heavy - OOP is not required. Apart from event handlers.
If you don't want to write in OpenCL to speed up calculations, fine, don't do it. The code will still run, although maybe not as quickly. If you don't want to use the Cloud, don't. But then make sure that your optimization (if you will use it) didn't take so much time. These are all tools to speed up your development and execution. They are not mandatory to use.
Anyway, the main thing, without which you will never get a profit, is the system itself. If there will be a system - all other issues will only become technical.
P.S. about ninja/dukas: also - look at their services. Do they have anything similar to our "Work"? As far as I understood, the forum of both are full of trivial questions for programmers. Ask for prices of their services.
Technically, Dukas has an MQL4 -> J++ converter, but it's not very well made. You still need the help of an experienced coder, and it may cost you a pretty penny.
So think about who those platforms are for (traders or coders) and who is for MT5.
I, on the other hand, work in DOS. It's all here http://old-dos.ru/
It's got drivers for all kinds of peripherals (CD-DVD, flash, sound, video, etc). Everything flies. What are these new monsters for? It's a nuisance.
I don't know how much. Enough so that modifying it after a couple of months or six months doesn't turn into an almost impossible task, which is as difficult as writing all the code again.
Well, they used to write operating systems without OOP. And nothing, the operating systems worked.
P.S. Now I'm not even talking about OOP, but about data types. You will feel the edge of necessity of switching to a five by yourselves - for example, as soon as you get tired of creating numerous "parallel to each other" arrays in a four, imitating a "structure array" data type. Or have to process a list/tree with addition/deletion of elements.
Still, the main thing without which you will never get a profit is the system itself. If there is a system, all other issues will only become technical.
The point is that there is a system. It is not super-profitable, but it exists. And there is only one question, and it is a technical one: how to increase the speed of optimization for this very system?
The answers are there too:
1. Increase the hardware
2. See item. 1, and don't you dare to claim ("blatantly bullshit" as Renat put it) that the platform is hardware-independent.
Renat:
......
This is part of what has been done for regular traders.Indeed, the maintenance of the MT5 product is several orders of magnitude higher than in MT4, a person who wants to use ready-made solutions has all the conditions, and what is important - all the services are in Russian
The fact that mql5 syntax is similar to C++ creates the illusion of similarity of programming languages, and porting simple codes from C++ to mql5 makes you think about the feasibility of porting codes with trivial connection of .dll created in the same C++, imho
ЗЫ: huge computational potential of МТ5 makes me wonder "why we need such computational power in terminal/tester?", Open Source neural network project seems to be "stalled", do we really need such speed to compute/optimize MAUs?
IgorM:
ZS: the huge computational potential of MT5 makes one wonder "why do we need such computing power in the terminal/tester?", the Open Source neural network project seems to be "stalled", do we really need such speed to compute/optimize MAUs?
Of course we do! What's unclear here? Otherwise, how to fit into the constraints of the Clouds Network? Other way is not possible, because if the number of input parameters exceeds the norm, the genetic algorithm will switch on and spoil the whole Scheme.
But what a rush! You'll tell later in retirement how the distributed network has grinded through a brute force attack on thousands of computers the input parameters of mashka's TS in less than a minute.
And if you also buy a dozen of other video cards and rewrite Mashka's Expert Advisor for OpenCL, the buzz will increase exponentially, and you'll be pissing without a break right and left, like a fire hose in honor of the technical revolution.
It does not matter that the TC is unlikely to be profitable. The main thing is to participate, not to win.
So quit MT4 and switch to MT5 to keep up with the technological progress.
ZS: huge computational potential of MT5 makes one wonder "why do we need such computational power in terminal/tester?", Open Source neural network project seems to be "stalled", do we really need such speed for calculation/optimization of MAUs?
With computational power, it is possible to solve "head-on" tasks that previously could have been abandoned due to impossibility/complexity of implementation.
For example, a simple conversion to 64 bits has given us a lot of advantages in our work. In one of our projects we enabled the in-memory cache up to 20 gigs (it depends on the available memory) and got an order of magnitude increase of the re-indexing speed of huge databases. Many of our customers all over the world use this feature now. The same can be done with MQL5 Cloud Network or by simply increasing the number of local agents in your network.
The Open Source project has not stalled, but we have had to delay using the native MQL5 Storage implementation in the editor. Without this, collaborative work is almost impossible. It will be out in the next couple of weeks already.
So dump MT4 and switch to MT5 to keep up with technological progress.
By the way, progress is coming like a steam locomotive and not much attention is being paid to anyone.
Look at how much computer power walks the streets in people's pockets and you'll find another confirmation that the world has gone crazy a long time ago.
With computing power, it is possible to solve problems "head-on" that could previously have been abandoned due to impossibility/complexity of implementation.
For example, simply moving to 64 bits has given us a lot of advantages in our work.
Renat, that's not what we're talking about. Stupid traders ask what they will get out of technological progress? What it gave you, they are not very interested. So they are speculators, selfish and mercantile.
However, I will not interfere. Go on; your speeches are quite charismatic and you make me run to buy a video card and a book on OpenCL.