Errors, bugs, questions - page 588

New comment
 
Rosh:
Have you tried ZeroMemory() ?

Applied....thank you....helped....)

 

2011.12.01 22:18:58 Core 1 'M.ex5' has newer unsupported version, please update your tester agent

How do I do what the terminal asks?

If you can't, why does it show this message?


 

I already addressed my question to SD, 10 days ago.

Roboforexe has 2 types of accounts, Instant Execution and Market Execution.In the first case the fill type is AON or FOK.In the second you have to use IOC.This is the experienced way.

In the beginning I tried to set the fill type automatically by 

int filling=(int)SymbolInfoInteger(_Symbol,SYMBOL_FILLING_MODE);
send.SetTypeFilling(filling);

I wrote a check script and asked a question to the BOD.


I have a problem with this script, but in practice it requires different values.

Maybe this is a bug on RoboForex server...

Files:
fill_script.mq5  1 kb
 

What's the bug with the indicators? They appear and then disappear. Only the ones that are in a separate window!!!

Here's a screenshot of when the indicators disappeared. They disappear now and then appear... arbitrarily. There's a video too...

Attention, the basic indicators are disappearing!!! That means the bug is significant. There's the same problem with custom indicators.

Gentlemen of the developers, fix this bug, it's not very nice...

 
555 tester. When trying to run a single test through the cloud (using local agents disabled, cloud agents enabled and in Ready state) the message "none of selected agents is ready, testing not started" comes up. Is this the intention?
 
Ashes:
555 tester. When trying to run a single test through the cloud (using local agents disabled, cloud agents enabled and in Ready state) the message "none of selected agents is ready, testing not started" comes up. Is this the way it's supposed to be?
Testing (not optimization) is only possible on local agents.
 
Please advise, in my EA, the indicator updates the value in the tester only with a new bar, that is, if the expert is running on the hourly chart - it will be updated once an hour, on the minute - once a minute. The Expert Advisor works through the onTimer function, the timer is set for 10 seconds. What may be the problem? I want it to update by timer. The iMA indicator
 
The logic behind this limitation is unclear.
Suppose I have a computer with a PR of about 10, and I want to run a test (the result without detail is enough for me) on a sufficiently large portion of history. Such a run would take several hours in these circumstances. In this situation it may be preferable to run through the cloud, moreover - the client pays.
 

Tester 555.

Optimisation through the cloud (100 passes) did not go all the way. Stopped at 97. Released after an hour of inactivity. Several passes with errors "no memory" or "cannot initialize expert", about 100 (almost all!) "optimization pass XX returned to queue", several "pass XX returned result 0.00 in YYYY ms (PR ZZZZ)" (not included in graph and optimization results!):


...

RO 0 MQL5 Cloud Europe 15:06:39 pass 37 returned result 9006.50 in 11 sec (PR 99)

HF 0 MQL5 Cloud Europe 15:06:40 pass 70 returned result 9006.50 in 43 sec (PR 125)

LM 0 MQL5 Cloud Europe 15:06:43 pass 82 with result 9006.50 returned in 4087 ms already processed by another agent

CQ 0 MQL5 Cloud Europe 15:06:45 pass 48 with result 9006.50 returned in 5 sec already processed by other agent

EL 0 MQL5 Cloud Europe 15:06:49 pass 24 with result 9006.50 returned in 10 sec already processed by another agent

HP 0 MQL5 Cloud Europe 15:07:04 pass 44 returned result 9006.50 in 3915 ms (PR 99)

QG 0 MQL5 Cloud Europe 15:07:20 pass 90 returned result 9006.50 in 5 sec (PR 59)

KS 0 MQL5 Cloud Europe 15:07:28 pass 69 returned result 9006.50 in 5 sec (PR 118)

FE 0 Tester 16:23:35 optimization completed, total passes 97 (successfull 87 passes)

HD 0 Statistics 16:23:35 optimization passed in 1 hours 42 minutes 41 seconds

DP 0 Statistics 16:23:35 locals 0 tasks (0%), remote 0 tasks (0%), cloud 120 tasks (100%)

HE 0 MQL5 Cloud Europe 16:23:35 connection closed

LN 0 MQL5 Cloud Europe 2 16:23:35 connection closed

QG 0 MQL5 Cloud Hong Kong 16:23:35 connection closed

QN 0 MQL5 Cloud USA 16:23:35 connection closed

QI 0 Tester 16:23:35 file cache C:\Program Files\MetaTrader 5\tester\cache\fourth_20110919_2_multi_SE.GBPUSD.M1.0.xml written

ID 3 Tester 16:23:35 stopped by user

Note that PR 59 (pass 90) returns result in 5 seconds, while PR 125 (pass 70) returns result in 43...

The application to the SR is being processed.

 
papaklass:
This is due to the fact that in the tester a single run is run in a single thread. Since a single run is a sequential process. That is, each subsequent result of this process depends on the result of the previous one. Parallelization of a sequential process is inexpedient from the technical viewpoint. It is profitable to connect the cloud when the processes are independent and executed in different threads. For example, optimization. Each run is individual and doesn't depend on the results of previous runs. That's why the optimization process connects to the cloud where each agent performs its own single run. I hope I've explained it clearly.

You have to look outside the box (c)

I don't see any contraindications to running a separate cloud run (except price).

1...581582583584585586587588589590591592593594595...3189
New comment
Reason: