Frequently Asked Questions

How does MQL5 Cloud Network work?

MQL5 Cloud Network is a distributed computing cloud network. Computing tasks are distributed to multiple computers registered in the network.

If someone needs to perform heavy computations which can be parallelized, the user can distribute the task to available processors. More about technical details.

Can I join the network?

MQL5 Cloud Network is free to join: download MetaTrader 5 Strategy Tester Agent and install tester agents on your computer.

How do I register?

Open the registration form and fill in the required fields. After that you will receive a registration confirmation email.

Please note that you can withdraw money via PayPal only to the email specified during registration. This protects your money from unauthorized access.

How to create a distributed computing task?

A task is a program in MQL5. Usually, distributed computing is used when someone needs to obtain many results for the same algorithm with different initial data. The Strategy Tester in the MetaTrader 5 client terminal supports such multi-threaded calculations for trading strategy optimization tasks.

Do I need to install the MetaTrader 5 terminal?

No. If you are only going to provide your computer's spare processing power, simply download MetaTrader 5 Strategy Tester Agent and install tester agents on each computer that you want to connect to the network.

You may need the MetaTrader 5 terminal, if you want to run MQL5 applications containing a trading strategy or a mathematical calculation task.

Which programming language does the MQL5 Cloud Network use?

It uses the MQL5 trading strategy programming language. The language is very close to C++ in terms of syntax and execution speed. Therefore, you can quickly learn MQL5 and convert algorithms from other high-level languages.

The MQL5 language is part of the MetaTrader 5 trading platform. A plethora of comprehensive educational materials are available on the community website.

Is providing processing power via the MQL5 Cloud Network safe?

It is absolutely safe. Each tester agent runs in its own virtual sandbox, which prohibits external requests from MQL5 programs. The sandbox blocks any DLL calls and access to files outside the specified directory.

Can anyone get my program code or other data if I use MQL5 Cloud Network agents?

No. Data exchange between the computations initiator and the tester agents is encrypted. The tester agent does not save the received program code on a drive, and thus it cannot be intercepted on the computer. Tester agents are protected from any attempts to intrude into their runtime environment and capture data from the agent's memory.

I installed agents and entered my MQL5.com account credentials, but I don't see the agents in my Profile.

The agents will appear in your Profile after completing the first task.

The number of agents displayed in my Profile has been reduced. Why?

You might have enabled the "Active last day" filter. Uncheck the box to see all your agents.

How do I find out my agents' PR?

Each agent's PR is displayed in the owner's Profile. The value may vary depending on the computer load.

I select slow optimization (to test all parameters) for my task, however it runs as fast optimization using a genetic algorithm. Why?

If the total number of optimization passes exceeds 100,000,000, fast optimization is selected automatically. Read more.

Why is optimization not available on 32-bit systems?

We discontinued support for 32-bit systems because they do not meet modern requirements for distributed computing, as well as reduce performance and add overhead when maintaining the network infrastructure.

Complete optimization uses all servers and free MQL5 Cloud Network agents, while genetic optimization only uses the agents of one node of the MQL5 Cloud network. Why?

With genetic optimization, the set of all passes is replaced by a combination of genes which generate a limited set of passes (no more than 256) per population. The genetic algorithm waits for the computation of all passes within one population to complete before processing their results and creating new passes in the next population.

Therefore, the most reliable way is to distribute genetic optimization tasks to the agents of one MQL5 Cloud Network access points. Read more.