All (not yet) about Strategy Tester, Optimization and Cloud - page 18

 

The question:

Forum on trading, automated trading systems and testing of trading strategies

New MetaTrader 5 platform version build 5200: OpenBLAS extension and increased control in MQL5

leachimk , 2025.08.03 12:10

My computer has 32 logical cores. After updating the strategy tester, there are only 16 of them left. Is it possible to increase them to 32?

The reply:

Forum on trading, automated trading systems and testing of trading strategies

New MetaTrader 5 platform version build 5200: OpenBLAS extension and increased control in MQL5

Renat Fatkhullin , 2025.08.03 15:50

For internal tests inside the terminal, you have all ... cores shown, except for Efficient Low Power. LP cores (not to be confused with multi-core) are practically useless and are not even used by the operating system in normal mode. LP cores simply mislead people on such hybrid processors as Intel Core Ultra 185H :

...

But in MQL5 Cloud we only accept physical cores (...). LP cores are not physical and are always ignored.

---------------------

The question:

Forum on trading, automated trading systems and testing of trading strategies

Agents and Ryzen Ai 9 365 not detecting all cores and threads

Andres Lume , 2025.08.31 21:35

In my opinion , Low Power does the job well. Slower, but it does the job. So what's the difference? The work will be done faster than without them anyway. Out of 100 passes, they will do 10-20, but they will do it. 3.3 mhz speed of my zen5c, this is the level of normal processors 3-4 years ago.

I just don't see the point in blocking any agents, even the slowest ones. At least for local work.

Is it still not possible to connect phones as agents?


The reply:

Forum on trading, automated trading systems and testing of trading strategies

Agents and Ryzen Ai 9 365 not detecting all cores and threads

Renat Fatkhullin , 2025.09.01 08:08

You are confusing Low Power(LP) and Efficiency(E) cores.

LP cores are not even used by the operating system in normal operation. They are intended for use in the laptop's off/sleep/semi-sleep mode.

With 5c cores there is an error in identifying E cores as LP. We will fix this.

Normal cores are P(erformance). Intel has seriously messed up the architecture with P, E and LP cores. Now, to write efficient programs, every programmer must clearly track the core types, monitor which task to assign to which core and clearly fight against E and LP cores


and -

Forum on trading, automated trading systems and testing of trading strategies

Agents and Ryzen Ai 9 365 not detecting all cores and threads

Renat Fatkhullin , 2025.09.01 08:23

Advice for fans of synthetic tests:

  • if you have P/E cores, then you can't trust the tests
  • until you explicitly and firmly set Affinity Mask on the P core, you can't trust micro/nano tests
  • if the processor frequency is not locked (turbo boost and/or C/P-states are enabled), micro/nano tests cannot be trusted
  • if you have a mid-generation Intel (not late 2024/2025) and the patches (they are installed automatically) on the processor are not disabled, you cannot trust the tests
  • if you have Intel Xeon under clouds (AWS, etc), just throw away the tests - there is such a layer of applied patches that it is impossible to talk about performance at all
  • It makes almost no sense to conduct tests under clouds (even your own private ones)

 

Example - 

============

The question:

Forum on trading, automated trading systems and testing trading strategies

MT5 issues after update: CPU agent(core) reduced from 16 to 8

Máté Rémiás, 2025.08.08 08:48


Thanks Miguel for the quick reply, I have read all the comments in theis and referenced threads and my case is still not answered hence I posted here. Let me go in details:

- I have a 14700k which has 8 P cores, 8 HT threads and 12 E cores.

- MT5 Build 5200 (linux) only sees 16 core, earlier versions showed all 28 as possible agents.

- I use only local agents no local network farms nor cloud network.


I would expect to be able to decide how many of my cores to use for optimization on my own, even if HT cores are slower.

if I take the post you referenced I should see all my cores in Local, which is not the case. in monitoring it displays (see my 2nd attached image on the previous post) that 8 P cores are used (maybe HT cores are skipped), but out of my 12 available E cores only 8 are used, seemingly with the HT threads being not available as well, so the explanation doesn't add up.


The reply:

Forum on trading, automated trading systems and testing trading strategies

MT5 issues after update: CPU agent(core) reduced from 16 to 8

Eleni Anna Branou, 2025.08.08 13:17

You need to use the 'Local Network Farm'.

After the latest 5200 MT5 build update, in order to take advantage of all your processor's cores and threads in the Strategy Tester (locally), go to MT5 >> View >> Strategy Tester >> Agents >> Local Network Farm >> right click >> + Add >> Add Agents from Host (automatically) >> Next >> Done.



 

From Novice to Expert: Mastering Detailed Trading Reports with Reporting EA

From Novice to Expert: Mastering Detailed Trading Reports with Reporting EA

Today’s discussion focuses on addressing the challenges associated with delivering trading reports in MetaTrader 5. In our previous publication, we outlined the overall workflow and requirements for the system to function effectively. We introduced the Reporting EA as a tool designed to generate and deliver trading reports in PDF format, with a frequency that can be customized by the user.
From Novice to Expert: Mastering Detailed Trading Reports with Reporting EA
From Novice to Expert: Mastering Detailed Trading Reports with Reporting EA
  • 2025.08.28
  • www.mql5.com
In this article, we delve into enhancing the details of trading reports and delivering the final document via email in PDF format. This marks a progression from our previous work, as we continue exploring how to harness the power of MQL5 and Python to generate and schedule trading reports in the most convenient and professional formats. Join us in this discussion to learn more about optimizing trading report generation within the MQL5 ecosystem.
 

Post-Factum trading analysis: Selecting trailing stops and new stop levels in the strategy tester

Post-Factum trading analysis: Selecting trailing stops and new stop levels in the strategy tester

In the previous article, we created an EA that trades in the client terminal's strategy tester based on the results of trading on a real account. We have added the ability to set new StopLoss and TakeProfit sizes to test our trading in the tester with different stop order sizes. The result was unexpected: instead of a loss, we received a profit commensurate with the loss incurred in real trading. This means that if we had used the StopLoss and TakeProfit levels in our trading account that allowed us to make a profit in the tester, then trading on a real account would also be profitable. And this is just a simple change in the stop sizes. I wonder what would happen if we added a trailing StopLoss? How would this change things?
Post-Factum trading analysis: Selecting trailing stops and new stop levels in the strategy tester
Post-Factum trading analysis: Selecting trailing stops and new stop levels in the strategy tester
  • 2025.10.06
  • www.mql5.com
We continue the topic of analyzing completed deals in the strategy tester to improve the quality of trading. Let's see how using different trailing stops can change our existing trading results.