My approach. The core is the engine. - page 40

 
Maxim Kuznetsov:

and this is a free and affordable product? where have you seen this...where have you seen the product in question at all :-)

Peter doesn't hide the fact that the code will be closed, and he hesitates whether it will be paid and on what basis.

Once again: you can discuss closed source code. You can even write about paid products in the Marketplace. If you don't believe me take a look at my profile and read the rules again. What Peter is going to do with his code is his own business. Intent is not a presumption of guilt.

We're all here to pick Peter's brain. It's useful and constructive. So please don't discredit this thread. There's already a lot of flubbing going on.

 
TheXpert:

here. in general, it would be good to attach this post to the start post somehow, so that there would be fewer such questions

It's fine. Engine with closed code? So really TC is not a crank, but a clever advertiser who has successfully collected 39 pages of comments in a few days. I don't know if it's allowed by the forum rules, but everyone posting here plays into his hands.

 
Georgiy Merts:

How does that invalidate what I said?

C code tends to be close to assembly code, but it is still assembly code that has the greatest efficiency. Whatever the year.

I don't understand about "loading CPU cache by gauge method". That's why the assembler is used to load the cache as effectively as possible! What gauge method?

No! Just because the C++ syntax means nothing

With your approach to modern processors only on the forums arduinchiki programmer - there as you write, here's a compiler C, it kick everything into Assembly=machine code and want to directly Assembler write. if you want to write directly in a machine code in memory the commands of the processor - the result will be identical, because there is a single-chip processor with the usual aparticle registers and direct memory access, no virtualization, cache and instruction pipeline - that is in modern processors and main advantage, which allows you to add performance to new processors at already achieved maximum processor clock speeds

 
Igor Makanu:

Hmmm, I didn't think I was going to argue with you, have you looked at the calendar lately? What year is it? What inefficiencies in interface gadgets? You mean the memory organization...

Don't make such a big deal out of it. What Georges meant is clear to everyone. If you want program speed - program it directly. If you want simplicity - use different layers. That's what I was talking about.

 
Vasiliy Sokolov:

Stop making such a big deal out of it. We all know what Georges meant. If you want program speed, program directly. If you want simplicity, use different layers. That's what I was talking about.

I'm not making a hump nor a limp, above@Ilya Malev he wrote, that he didn't like that in MQL an object occupies 24 bytes, everything looks logical, there's much room for the structure, but it's in those moments and hides the optimal work with the PC hardware, where with caches, where with CPU registers, where is the preselection of commands, and where are predictions of transitions and branches, but just trying to arrange everything compactly by types may lead to loss of productivity

imho, there are well-established tendencies in programming - OOP, work with databases, it means that at the moment these are the most optimal solutions, both hardware manufacturers and compiler manufacturers adjust to them and there is no need to build a dummy ))))

HH: I'm not even going to search for old compilers, but I suspect that if we test the well-known compiler from the region of 2000 and its latest versions, on a completely identical test case, we'll see the difference in speed, since the new compiler will be optimized for new generations of processors. I also think that if assembly inserts are available in the compiler then it will be impossible to find your assembly code in the compiled project under the Asembler debugger, because the compiler will only interpret the work of the assembly insert.

 
Vasiliy Sokolov:

What are you talking about? What commercial product? Where is the link to it? Is it available on the Market? Compiled ex4 can be posted and discussed if anything.

Page 30.

Where did you read that ex4 can be posted and discussed?
 
Yury Kulikov:

No, the "kernel" of the topicstarter is an array of strings of immense size, and it is unrealistic, even theoretically, to talk about efficiency of such an approach.

The core is an array of int type. It is two-dimensional. Size = 236 properties*number of objects.

1 button = 3 objects (base, text, icon).

Most elements are 1 - 2 objects.

Execute list - 11 objects.

Tab - 4 objects.

Slider - 4 objects.

"Unobstructed" is a strong word)).

 
Реter Konow:

The core is an array of int type. Two-dimensional. Size = 236 properties*number of objects.

1 button = 3 objects (base, text, icon).

Most elements are 1 - 2 objects.

Execute list - 11 objects.

Tab - 4 objects.

Slider - 4 objects.

"Unobstructed" is a strong word)))

This is progress :))

It was different in the past, you may recall the gradient function.

Forum on trading, automated trading systems and trading strategies testing

Colour decomposition function.

Retag Konow, 2018.10.11 12:43

About the difference in speed of the algorithms:

1. Your algorithm fills uint array and mine fills string array. (I need exactly string, to easily extract the right components from the hue and then work with them).

 
Vasiliy Sokolov:

In short, you guys are wrong to keep cornering Peter and in the process leading the thread into flub.

Peter has posted something of an engine. This can be compiled and run. Then hopefully constructive discussion will start: discussion of the engine, its interfaces and how to do this and that.

Get out of the corner and become a developer.

Do you propose to give in to this kindergarten? After all, more than once expressed constructively about it. But every time, with increasing frequency, Peter is dumping his gui on us. A clear promotion of his commercial project. He's not even hiding it :)

 

I don't know how it is.

but I've dreamt about this before - the core engine

Oh, man.

I have to go back to the nuthouse for a rest.