MQL5 Cloud Network: Are You Still Calculating? - page 8

 
I'm glad that you are moving towards GPU computing, because my 16x7970s have been waiting for a long time to join the cloud) But will there be so much work for them?
 
Tell me, does the agents work continue in the background when closing the agent manager, or does it stop?
 
sovetnikmaker:
Tell me, does the work of agents continue in the background when closing the Agent Manager, or does it stop?
It continues. The manager - "manager", gives a command to switch on, then the agents themselves, the manager is free.
 

I have 2 computers available to the cloud....... they are not used a lot, perhaps because of my location.


I quite like the idea of having other computers available in case I need results quickly, and the costs are trivial compared to the costs of having additional computers and maintaining whatever may be necessary to have them available.


This is a great service.!!

 
Статистика - MQL5 Cloud Network
Статистика - MQL5 Cloud Network
  • cloud.mql5.com
Произведенный объем работы, распределение агентов по странам, RAM, CPU и битности OS
 

Agents are not doing calculations (and are invisible in the profile at https://www.mql5.com/ under Agents)...

In the log:

EH 0 00:47:27.930 Startup access rights to common directory successfully checked

DQ 0 00:47:27.930 Startup Service start initialized

RF 0 00:47:27.930 Startup Create startup thread

MR 0 00:47:27.930 Startup Thread successfully created

OK 0 00:47:27.930 Startup MetaTester 5 x64 build 881 (11 Dec 2013)

JQ 0 00:47:27.930 Server MetaTester 5 started on 0.0.0.0:2000

OE 2 00:47:27.930 Agent OpenCL initialization skipped due to critical error, please update OpenCL drivers

HH 0 00:47:27.930 Startup initialization finished

Dedicated ip address is confusing... Internet is configured on the modem, port forwarding is configured 2000-2001 via tcp/udp protocols.

 
LiveWizard:

Agents are not doing calculations (and are invisible in the profile at MQL5 Cloud Network it is not necessary to forward ports, because in this case the agent works as a client and connects to the pool itself, but does not accept connections from outside.

Is this the full log of the agent? Is nothing else written?

Распределенные вычисления в сети MQL5 Cloud Network
Распределенные вычисления в сети MQL5 Cloud Network
  • cloud.mql5.com
Заработать деньги, продавая мощности своего компьютера для сети распределенных вычислений MQL5 Cloud Network
 

No more problems. Port forwarding has been removed. Agents perform calculations and information about them appeared in the profile.
This was only part of the log, but it doesn't matter....

Now I noticed a strange activity of agents (and remembered that I had observed the same thing before).

Periodically (1-5 times a minute) the first and second agents, and sometimes they simultaneously, within a few seconds significantly (up to 30-35%) load the processor, otherwise not loading it at all (0% load in Task Manager).

At the same time, the last entries in the logs are dated more than an hour ago:

ER 0 11:15:13.155 Network connected to 1.agents.mql5.com

GM 0 11:15:43.143 Network connected to 1.agents.mql5.com

PD 0 11:16:13.151 Network connected to 1.agents.mql5.com

QO 0 11:16:43.153 Network connected to 1.agents.mql5.com

CG 0 11:17:13.144 Network connected to 1.agents.mql5.com

QO 0 11:13:01.462 Network connected to 3.agents.mql5.com

DG 0 11:13:31.462 Network connected to 3.agents.mql5.com

EL 0 11:14:01.470 Network connected to 3.agents.mql5.com

OE 0 11:14:31.471 Network connected to 3.agents.mql5.com

LR 0 11:15:01.472 Network connected to 3.agents.mql5.com

DJ 0 11:15:31.464 Network connected to 3.agents.mql5.com

ES 0 11:16:01.472 Network connected to 3.agents.mql5.com

JH 0 11:16:31.469 Network connected to 3.agents.mql5.com 

And it's 12:44 on the clock.

According to the same logs, the last calculation was finished ~2.5 hours ago.

What is the reason for such behaviour of agents?

UPD: in the agent manager the number of passed tests grows. The number of passes also grows in the profile.

So it's the agents doing the calculations? But why is their behaviour so atypical - they do not constantly load the processor, but only "intermittently", periodically and (in my opinion) too rarely (as a result processor power is used inefficiently); RAM is not loaded (for example, now both agents use 12040KB, but usually during testing the amount of memory used grows significantly (by tens of MB)).

Распределенные вычисления в сети MQL5 Cloud Network
Распределенные вычисления в сети MQL5 Cloud Network
  • cloud.mql5.com
Заработать деньги, продавая мощности своего компьютера для сети распределенных вычислений MQL5 Cloud Network
 
Часто в расчет попадают задачи на маленьком интервале времени - поэтому такие задачи будут потреблять мало памяти и очень быстро заканчиваться. 
 
Hello everyone! I would like to those who provide processor capacity would pay more, well, or there are bonuses added for every 100, 200 thousand passes. And I would also like to appeal to the developers, maybe they will read this article to add to the user statistics number of passes performed for the current day, now displayed only total for all agents
Общайтесь с разработчиками через Сервисдеск!
Общайтесь с разработчиками через Сервисдеск!
  • www.mql5.com
Ваше сообщение сразу станет доступно нашим отделам тестирования, технической поддержки и разработчикам торговой платформы.