MT4-Tester VS MT5-Tester - page 7
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
Thanks - but I was talking about MQL4, including the braking there - apparently I didn't explicitly mention it.
ZZY In the meantime, one bonus from the comparison is almost immediately available. Looks like another one is maturing.
No more bugs could be found. The testers settings are different. For example, maximum number of open positions, etc.
Therefore, we had to tweak the code considering the differences to make the results coincide perfectly.
After the edit, optimization on MT4 took 13:50, MT5 (single core) took 25:02. The difference is 1.8 times in favour of MT4.
I guess the developers will tweak something else to reduce this gap.
At the moment MT4+TDS allows making custom history only using a crutch - emulate LZMA storage format of Dukascopy. But in future builds the custom history will be definitely available. Let's hope that on MT5 this event is not far off.
To summarize, I would like to say that MT4->MT5 and MT5-SB->MT4 converters have shown themselves without any failures. And that the tester settings can be much wider if desired.
Why don't you quote the time using all cores in MT5?
Isn't it obvious? To compare them on a level playing field.
No one is saying that MT5 is slowerin the complex. What is being compared is what can be compared - a single pass on a single core.
Isn't it obvious? To compare them on a level playing field.
No one is saying that MT5 is slowerin the mix. What is being compared is what can be compared - one pass on one core.
Everything is obvious to me.
But conclusions are drawn quite unambiguously and as far away as possible from even the thought of multithreading. If multi-threaded results were specified side by side as I did, I wouldn't have any complaints.
The main point of this and similar threads is to imagine that MT5 is slower in testing. As if it's my first day in the world and I don't see a bunch of attempts on masses of sites to compare in platforms only what one wants by the lowest denominator in a mode of complete silencing of 80% of the features.
It's obvious to me.
But the conclusions are completely unambiguous and distance themselves as much as possible from even the idea of multithreading. If multithreading results were indicated next to each other, as I did, there would be no complaints.
The main point of this and similar threads is to present that MT5 is slower in testing.
I haven't seen any claims, just attempts to sort it out. And I see the aim of the thread as different - to improve MT5.
Part of it has already been achieved, right here a new build has been assembled and posted for testing, which is times faster in certain conditions.
I haven't seen any complaints, just attempts to sort things out. And I see the aim of the branch as different - to improve MT5.
Part of it has already been achieved, right here a new build has been built and posted for testing, which is several times faster in certain conditions.
The build has been built before, just hasn't released the beta yet.
Any meaningful public text needs to be analysed not on the scale of the 10 participants who fully understand the discussion, but on the scale of the chunky perception of the masses remembering the torn out statements. This is why I have to participate in the masses of discussion, correcting the situation.
Why don't you quote the time when using all kernels in MT5?
The answer above is correct. It was the testers I was comparing, i.e. single run. And optimizer used only as a tool to make more statistical sense.
Today for the first time I've had the opportunity to compare two testers. It's a pity the trial will end in a fortnight. After the tests I started to have more confidence in MT5, because even though it was a simple test, it showed coincidences.
I think it's a good tone to check your products for correctness by comparing with alternative solutions.
Frankly speaking, I liked working with MT4 tester better, especially concerning GUI, because it's the first time I've used it for a long time, while I used MT5 almost every day.
And I know in MT4, even without TDS, how to speed up my Expert Advisors through custom history without any loss of accuracy. In MT5, unfortunately, it cannot be done yet.
I made TDS leave duplicate ticks for the sake of the experiment. But I have not lost the accuracy, I could have filtered them using TDS tools. And the speed would increase 2 times more. I would like the MT5 tester would not create duplicates on FOREX symbols. Ideally, of course, it should be custom.
I hope I am not the only one who found this thread useful.
If multi-threaded results were indicated next to each other, as I did, there would be no complaints.
What is the point of this - single MT5 pass uses a single core, while MT4 optimisation runs a terminal for each core.
I have to be honest, I was much more pleasant to work with in the MT4 tester, as far as the GUI is concerned, even though it was the first time I ran it in a long time, while MT5 was used almost every day.
100% true statement!