Errors, bugs, questions - page 2610
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
Then you can't go through several advisers. Thank you.
MT5, both options.
Googled the question, in case I missed something, but everywhere the answer to the question "run a thread on a physical core" is that Windows Task Scheduler does not allow to do this
i.e. if it's principled to solve a task in several threads (not physical cores!) then only launch multiple copies of terminal, but i think it will make sense up to certain CPU load, then task scheduler "deal with this nonsense" based on its vision of task priorities in Windows
Thanks, I'll do it through the Agents.
Thank you, I will do it through Agents.
I have managed to load all cores with 1 EA, I need 3 Gb of RAM.
I took the BitonicSort.mq5 script in the examples, changed OnStart to OnInit
Example attached
It managed to load all cores with 1 EA, it requires 3 Gb of RAM.
I took the BitonicSort.mq5 script in the examples, changed OnStart to OnInit
Example attached
ran your code in my environment
the terminal had 19 threads in Windows Task Manager. After running your code, only one thread was added to Task Manager in the "threads" column
How did you check?
but in my opinion up to a certain CPU load it will make sense, then the task scheduler "will deal with this mess" based on its vision of task priorities in Windows
as expected, on Win10 the task scheduler behaves differently, ran this code, the CPU is 100% loaded and the task scheduler started to block it
If running 5 EAs - will they run on one core or five? If on one, then the only way is Tester Agents?
Each EA works in its own thread.
Turn on the threads column in your task manager and see how many dozens of threads the terminal actually uses.
I managed to load all cores with 1 EA, it requires 3 Gb of RAM.
I took the BitonicSort.mq5 script in the examples, changed OnStart to OnInit
Example attached
Unfortunately, I am a total zero in OpenCL, but would like to learn. Today I will post a calculation script on the blog.
If you would like it, it would be great to demonstrate OpenCL capabilities on it. There are billions of independent subtasks in there.
Every expert works in their own stream.
Turn on the threads column in your task manager and see how many dozens of threads are actually used by the terminal.
Unfortunately, incompetent. Therefore not aware of whether these threads will be distributed simultaneously between different processor cores.
ran your code on my computer
the terminal had 19 threads in Windows task manager, after running your code only one task manager thread was added to the "threads" column
How did you check?
As expected, on Win10 the task scheduler behaves differently, here is running this code, the CPU is 100% loaded and the task scheduler started to block it
Opened 1 clean chart (without indicators). There was a question about parallel computing on all cores or one core.
As a result, the advisor can engage all cores, all 8 cores engaged.
Forum on trading, automated trading systems and trading strategies testing
Bugs, bugs, questions
fxsaber, 2019.12.03 14:35
Didn't find it by searching. Need to parallelize calculations. If run 5 EAs - will they run on one core or five? If on one, then the only way is Tester Agents?
There is 1 clean chart open (no indicators). There was a question about parallel computing on all cores or one core.
As a result, the Expert Advisor may engage all cores, all 8 cores are engaged.
You are probably right, I googled the threads-core topic, one program thread can run simultaneously in several cores
Well, I'm not sure about it.
Unfortunately, I'm a complete novice in OpenCL, but would like to learn. I will post a calculation script on the blog today.
If you would be willing, it would be great to demonstrate OpenCL capabilities on it. There are billions of independent subtasks in there.
https://www.mql5.com/ru/blogs/post/731774