Using the Cloud for GP Computing
Everything about Cloud is collected on the first post of this thread by links (thanks to angevoyageur) :
All (not yet) about Strategy Tester, Optimization and Cloud
Everything about Cloud is collected on the first post of this thread by links (thanks to angevoyageur) :
All (not yet) about Strategy Tester, Optimization and Cloud
Thank you for the information, but Does this answer my question?
Thank you for the information, but Does this answer my question?
Go to the link and see - if you will find some general answer by link so it will be fine ... because I am not fully understand your question sorry - especially "large array divided to multiple agents"
it is just one moment to read because it is one post only here
Go to the link and see - if you will find some general answer by link so it will be fine ... because I am not fully understand your question sorry - especially "large array divided to multiple agents"
it is just one moment to read because it is one post only here
I am sorry, I did go to the link and didn't find any answer.
I don't think there is an answer referenced in that topic.
When you run an optimization, each agent run the same code but with different parameters, so if you can implement your algorithm to follow this paradigm, the Cloud can be used to GP computing.
Can anyone tell me how do I make sure that the result I get from the cloud on 15 minutes chart of a whole year is the exact results I would get if I would perform it on my private machine?
How do I know that the charts on the clients of the cloud are the exact charts of the whole year on my machine?
Thank you for the answer.
Can anyone tell me how do I make sure that the result I get from the cloud on 15 minutes chart of a whole year is the exact results I would get if I would perform it on my private machine?
How do I know that the charts on the clients of the cloud are the exact charts of the whole year on my machine?
Thank you for the answer.
So, my idea would be start two instances of MT5, with the same EA and selected input parameters range, one for local optimization (the slow one) and other for the cloud (the fast one).
At least the result would lead us to be near 100% sure to answer your question.
PCWalker, good point, since to be 100% sure about that it would be nice have a control on genetic optimization algorithm (check box option, for example), like we have when we use MathSrand().
So, my idea would be start two instances of MT5, with the same EA and selected input parameters range, one for local optimization (the slow one) and other for the cloud (the fast one).
At least the result would lead us to be near 100% sure to answer your question.
Thank you for your answer, does the Metatrader platform select agents from the cloud that are on the same broker?
It seems that out of 2000-4000 agents on each of the region it is using only 100-200 agents. And I don't know why.
Thank you for your answer, does the Metatrader platform select agents from the cloud that are on the same broker?
It seems that out of 2000-4000 agents on each of the region it is using only 100-200 agents. And I don't know why.
PCWalker, you are welcome.
As far as I know this is not a problem, but a design decision. To explore more agents you must select Slow Complete Algorithm for optimization.
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
Does anyone know about how is it possible to use the MQL Cloud for general purpose computing ? ( Articles )
Like for example using an algorithm on large array divided to multiple agents and returning the results to the source requested.
Thank you for any replies and responses.