
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
It takes 1 + a half minutes to run two passes (2) with a data file that takes 12 secs to load normally.
The tester function also has nothing in it.Just the load.
So , there is no caching from what i can tell .
Since there's so much overhead in loading and dumping and loading and dumping the same file over and over
its more efficient to construct your own genetic algorithm , load the data ONCE and have at it with one core.
dissapointing
EDIT :
Also , with property tester_file , you have to compile , WAIT depending on how big your file is , close the IDE , close MT5 ,
go to the tester folder . Delete all files , reopen MT5 and test.
EDIT 2 :
if i manage to make this work i'll post the method here
I think your tester has memory leak issues on math calcs mode.
I shrank the test file to 800MB , i've got 4 cores.
On a normal chart it loads fine and properly (the file with the same code)
On math calcs mode i run out of memory.
I assume it needs to load it 4 times , right ? one per core so 3.2 GB on a 16GB machine.
Memory bloats over 5GB at some point.
I have a strict ZeroMemory instruction after the load too , no change.
The load function unloads all arrays before it loads the file
I checked my save/load functions for consistency.
I assume there is an issue with loading doubles on the tester from a file.
Will try to replicate with sharable code.
Cheers
test scripts .
Let me know thanks .
[_]p beer
PS : dont try it on your main pctest scripts .
Let me know thanks .
[_]p beer
PS : dont try it on your main pcIt seems you don't understand that a data file of 900 MB will create memory objects (you have 3 imbricated arrays of classes) using around 10 times more memory ?
Do you think that using structs, classes and arrays is not using any memory ?
I run it with 1000 shelves instead of 10,000 : data file is 94.8 MB while the Terminal is using 1165 MB to keep them in memory.
It seems you don't understand that a data file of 900 MB will create memory objects (you have 3 imbricated arrays of classes) using around 10 times more memory ?
Do you think that using structs, classes and arrays is not using any memory ?
I run it with 1000 shelves instead of 10,000 : data file is 94.8 MB while the Terminal is using 1165 MB to keep them in memory.
No i dont understand that thats true .
So the "tree" is the issue ?
I have to reduce dimensionality?
And i'll also have to take into account user's mem size in exports too if that's the case.