Optimisation in the Strategy Tester - page 6

 
Mr.FreeMan:
Hurray, the tester is running at normal speed now. If you remove the limitation on the number of optimized parameters, I may think about switching to the tester by MetaQuotes Software Corp :)

And what do you personally do not like the possibility to optimize for 64 parameters (this is only explicitly indicated)?

It's hard to imagine how many options you can get for optimization...

 
Interesting:

And what is it about the ability to optimise by 64 parameters that bothers you personally (this is only explicitly stated)?

It's scary to think how many variants you can get for optimization...

64? At 10 it says "Tester too many passes for optimization, please try to decrease optimized parameters amount or to increase parameters' values step".

at these parameters:

 
Mr.FreeMan:

64? At 10 it says "Tester too many passes for optimization, please try to decrease optimised parameters amount or to increase parameters' values step"

at these parameters:

According to the developers parameters can be guaranteed up to 64.

But here (in your case, as I understand it) there is another limitation - on the number of tester passes (though I do not remember the exact value).

The tester stated this directly to you (additionally highlighting the total number of passes in red).

 
When the number of variants is huge, use a genetic optimiser instead of a direct search. It allows efficient processing of trillions (any number) of variants.
 
Renat:
When the number of variants is huge, use a genetic optimiser instead of a direct search. It allows efficient processing of trillions(any number) of variants.
So it is with genetics enabled as well.
 
Mr.FreeMan:
So it's the same with genetics on.
We'll figure it out and fix it.
 
Renat:
We'll figure it out and fix it.
There have already been two or three upgrades, but the"64" limitation has not been removed. Now I will write one project and I would like to know exactly whether to use the standard tester with an unlimited number of runs or to adapt the code immediately to a self-made tester.
 
Mr.FreeMan:
I have already had two or three upgrades, and the"64" limitation has not been removed. Now I'm going to write one project and I want to know exactly, whether to use a standard tester with unlimited number of runs or immediately adapt my code to a self-made tester.

There is no error - maximum field of calculation should fit into unsigned int64 = 2^64-1(18,446,744,073,709,551,615).

When selecting the parameters to be run, use reasonable sufficiency.

 
Renat:

There is no error - the maximum miscalculation field should fit within unsigned int64 = 2^64-1(18,446,744,073,709,551,615).

Use reasonable sufficiency when selecting the parameters to be overridden.

I wonder what you mean by reasonable sufficiency?

Are you worried that my computer will consume a bit more power?

Why limit users artificially?

Another far-fetched problem.

 
Serj_Che:

Why limit users artificially?

More contrived problems.

OK, let's calculate. Suppose one pass in the tester takes 1 second, then let's divide number of passes 2^64-1 by 60 and get optimization time in minutes: 18 446 744 073 709 551 615/60.

Next, divide this time by 8 (8 cores on the computer) and get the optimization time with this number of passes on an 8-core computer. How many hours or days will that take?