New MetaTrader 5 platform build 4230: More built-in applications and expanded ONNX support - page 3

 
Vinicius Barenho Pereira #:

Hello Alexey I did the test with an EA from standard delivery and the problem persists.

When I click optimize using the local network farm the Metatrader client will go to 100% cpu use and stop responding.

I have a good cpu (i9 24cores +5ghz) and even if the option "Use Local Agents" is disabled the application will use 100% cpu and crash.

Please provide screenshots demonstrating it: tester with all the settings, agents, task manager with performance charts

 

New build 4231 has a BIG BUG. It cannot detect any objects on the charts on most VPS. EAs that use graphic interface get destroyed.

Please note that this issue usually happens on VPS, not on PC. Please check and fix it urgently.

 

Build 4231 Has a bug. Freezes and doesn't connect with farm Agents only couple random agents connects and then freeze. worked fine before. (i7-3630qm - 80cores)



Vinicius Barenho Pereira mine did the exact same thing worked fine before 4230

Timeline

* 09 march - I have work flawless (previous version mt5 release before 4230);
* 10 march - I have updated my local MT5 to version 4230 and noticed that after the update my local network stops responding and the reported error occurs;
* 10-march - new update release MT5 4231;
* I've try several times to run optimization without sucess. Before the crash I could see the message in the Journal asking my local network to update the client to the latest build. I can confirm my local network was updated to the latest build 4231.


 
Alexey Petrov #:

Please provide screenshots demonstrating it: tester with all the settings, agents, task manager with performance charts

Sure.

TEST 1 - Optimization of "Advisors\ExpertMACD.ex5" with local cores only. It works as expected

A) Settings
tester settings

B) Inputs

inputs


C) Agents using only local agents
local agents only

D) Optimization result ok

optimization result


TEST 2 - This one is using local network agents and the crash occurs, MT5 v.4231

A & B ) Settings and inputs are the same as previous test

C) Agents now use only local network farm
local network farm

Then I'll click on "Start". It will run, without giving optimization results, until the client stop responding.

D) Optimization crash



Timeline:

* 09 march - I have work flawless (previous version mt5 release before 4230);
* 10 march - I have updated my local MT5 to version 4230 and noticed that after the update my local network stops responding and the reported error occurs;
* 10-march - new update release MT5 4231;
* I've try several times to run optimization without success. Before the crash I could see the message in the Journal asking my local network to update the client to the latest build. I can confirm my local network has been updated to the latest build 4231.

 
Vinicius Barenho Pereira #:

Sure.

Provide your PC details please: exact CPU model, RAM, OS version

 
Vu Trung Kien #:

New build 4231 has a BIG BUG. It cannot detect any objects on the charts on most VPS. EAs that use graphic interface get destroyed.

Please note that this issue usually happens on VPS, not on PC. Please check and fix it urgently.

Too few details.

 
Alexey Petrov #:

Provide your PC details please: exact CPU model, RAM, OS version


My main PC specs are:

Processor: Intel(R) Core(TM) i9-14900KF
RAM: 32.0 GB DDR5 5600mhz
OS: Windows 11 Pro 23H2


My local network spec's
:

Processor's: i7-47900
RAM: 8 GB
OS: Windows 10 Pro


Thanks in advanced.

 

Hi,

Yes, same problem.

Build 4231 and 4232 always the problem. I expected amelioration on 4232 but nothing.

I have my main computer:

Processor: Intel(R) Core(TM) i9-13900 KS 6GHz (OC at 6.2 GHz) RAM: 196 GB DDR5 OS: Windows 11 Pro 23H2

Other computers with i7 and 64 GB RAM, another with i5... I also work with big servers with Intel Xeon Silver... same problem.

At the start, nothing worked; finally, I updated all the windows, server. The optimization works, but very very slowly on each agent. Some of them crash completely.

I have 138 agents in my farm; usually, I make for this type of job 5000/6000 tests on 24H (complex backtest on several years), and now only 250 tests in the last 10 years.

I have tried with half of the agents, same problem, only local agent with the i9 13900KS same problem. CPU used at 100% (usually 60/70%) and RAM used 180 GB RAM (usually 110). Of course for the same backtest.

I specify I have recompiled all my indicators, EA... to be sure to have the latest version.

Making optimization is my full-time job :( I'm at a complete standstill; I have to stop everything.

There are several of us, so there is a real problem with the latest versions.

Thank you

 
The bugs in the latest build, 4231 has rendered the GUI element of my EA virtually unusable. How do I restore the previous build until the next build is released? If this is not supported by the platform then can I request that it is implemented soon so we can at least have a temporary workaround. Thank you.
 

Hi,

also I have same problem, build 4231 and build 4232 beta.

Hardware:

Local: Processor: 2x Intel Xeon E5-2640v4 128 GB DDR4 ECC, OS Windows 10 22H2, 2 TB M2 NVMe

Using agents from: Processor: 1x Intel Xeon E5-2699v3  32 GB DDR4, OS Windows 10 22H2, 1 TB SSD


Thank you very much.