The /O3 effect is same with /Os which prefer size than speed

There is no relevance with "linker adjusts the address" in this case. The r8d register shown by the poster is an immediate value, not an address of a variable. The linker doesn't touch its operand at all. The only reasonable answer for that is because: 45 8d 41 0e lea r8d, [r9 + 14] is shorter than: 41 b8 0e 00 00 00 mov r8d, 14 The instruction lea r8d, [r9 + 14] adds the r9 value as the dependency. Thus mov r8d, 14 is better for the pipeline parallelism because it doesn't have any data dependency on the previous instruction. That being said: xor r9d, r9d lea r8d, [r9 + 14] is heavier for the pipeline, because the lea depends on the xor. So the lea can't be executed before the xor finishes. If you have: xor r9d, r9d mov r8d, 14 those two instructions (xor and mov) can be executed in parallel because the mov doesn't depend on the previous instruction. If you increase the optimization level to O3, it will choose mov over lea. See: https://godbolt.org/z/o85v97KoP

There is no relevance with "linker adjusts the address" in this case. The r8d register shown by the poster is an immediate value, not an address of a variable. The linker doesn't touch its operand at all. The only reasonable answer for that is because: 45 8d 41 0e lea r8d, [r9 + 14] is shorter than: 41 b8 0e 00 00 00 mov r8d, 14 The instruction lea r8d, [r9 + 14] adds the r9 value as the dependency. Thus mov r8d, 14 is better for the pipeline parallelism because it doesn't have any data dependency on the previous instruction. That being said: xor r9d, r9d lea r8d, [r9 + 14] is heavier for the pipeline, because the lea depends on the xor. So the lea can't be executed before the xor finishes. If you have: xor r9d, r9d mov r8d, 14 those two instructions (xor and mov) can be executed in parallel because the mov doesn't depend on the previous instruction. If you increase the optimization level to O3, it will choose mov over lea. See: https://godbolt.org/z/o85v97KoP

How to get it sent a huge message, it has been re-uploaded as a file # compile program #include <stdio.h> #define true 1 #define false 0 struct bTREE{ int height ; int left; int right;..

How to get it sent a huge message, it has been re-uploaded as a file # compile program #include <stdio.h> #define true 1 #define false 0 struct bTREE{ int height ; int left; int right;..

i want to reverse a string but my function is not giving any output

You need to resize your string s for the amount of elements

can anyone tell me what is the error

You are storing the left and the right subtree as an int and then casting them to a bTREE pointer. This is UB to begin with. And moreover on a 64 bit machine, ints will be 32 bits long and pointers will be 64 bits long. This would not even work. Why not work directly with bTREE pointers instead? Fix all these issues and then come back here if you still have problems.

why i can't send photos

why i can't send photos

The times.txt document is a 20-line file with hours:minutes:seconds per line. is a document. By reading this document, you create the always-existing data structure By assigning it to instances of its type, it will increment by 1 sec and print that increased time in a new document write the code. If the time format is misspelled, the user should be warned by skipping that line.

The times.txt document is a 20-line file with hours:minutes:seconds per line. is a document. By reading this document, you create the always-existing data structure By assigning it to instances of its type, it will increment by 1 sec and print that increased time in a new document write the code. If the time format is misspelled, the user should be warned by skipping that line.

The times.txt document is a 20-line file with hours:minutes:seconds per line. is a document. By reading this document, you create the always-existing data structure By assigning it to instances of its type, it will increment by 1 sec and print that increased time in a new document write the code. If the time format is misspelled, the user should be warned by skipping that line.

(int)1.0 vs int(1.0) what is the difference 🤔

(int)1.0 vs int(1.0) what is the difference 🤔

(int)1.0 vs int(1.0) what is the difference 🤔

int() is the constructor

typename class no difference la

Does C language utilize short circuit evaluation?

Who are you replying to?

typename class no difference la

There are minor differences between a struct and a class. int is definitely not a class. It is a built-in type.

Explicitly specify public/private stuff then everything the same la

Class? Built-in type? No difference leh

Yes there are. All built in type objects can be trivially constructed and copied. Not all class type instances can be trivially copied and constructed.

Can a class type be trivially copied and constructed?

If yes, then that’s all.

No

So you mean "no instances of class types can be trivially copied and constructed"?

I said only specifically designed ones can be trivially copied and constructed. For ex std::string is not trivially copyable. Built in types are specifically different from class types because the rules to handle them are specifically built into the compiler. The rules to handle classes are generic and not built in. Most of the optimizations that the compiler can do on a built in type are coded into the compiler (meaning the compiler can see a lot about them) while the ones that can be done on a class type instance are rule based learnings. So in effect built in types are different from compound types or user defined types. There is a reason why they are called built in types. That is the reason why no one will call int(1.0) a constructor call. It is called a functional cast. While on the other hand std::string("hello") would usually be considered as an invocation of a constructor .

Regarding optimization, there shouldn’t be any differences otherwise all the "zero-cost abstraction" effort will break. (Actually it already breaks but it’s a different story) From a users perspective, one can simply treat built-in types and custom types as the same. There is really no point distinguishing them.

Btw int is also a primitive in Java, where primitives and classes really have differences

<quote>--Regarding optimization, there shouldn’t be any differences otherwise all the "zero-cost abstraction" effort will break. (Actually it already breaks but it’s a different story)--</quote> I wasn't saying there is a difference in the optimization. I am saying there is a difference in the way it is carried out. And btw there are indeed optimizations that can be immediately visible on a built in type that may not be so for a User Defined Type. For example on an architecture where operations on an long long can be carried out atomically without a lock, a compiler would certainly be able to enforce this atomicity on all operations on long long objects. But on the same architecture if a User Defined type say A that measured out to be the same size as that of a long long, the compiler can't guarantee that operations on std::atomic<A> would be lock free. So yes there are some differences. <quote>--From a users perspective, one can simply treat built-in types and custom types as the same. There is really no point distinguishing them.--</quote> This I agree with on a superficial level.

The atomic example is an implementation thing, not related to optimization. Consider std::vector<bool> where it doesn’t make sense to say bool has anything special from that.

The atomic example is an implementation thing, not related to optimization. Consider std::vector<bool> where it doesn’t make sense to say bool has anything special from that.

The atomic example is an implementation thing, not related to optimization. Consider std::vector<bool> where it doesn’t make sense to say bool has anything special from that.

And I was always talking about the difference between built in types and user defined types from a compiler's perspective. So why suddenly are you changing track?

Class? Built-in type? No difference leh

Alright, different tracks all the way down 😂

Bruh

I was clear from the beginning on what I meant. You were just clearly confused on what was being said and went down a rabbit hole.

Well I admit I’m confused when you start to mention User-Defined Types

All class types are User Defined Types. Doesn't matter if it is explicitly defined by a user or by the standard library.

I just checked the reference that you are correct. I think next time you can just throw out std::is_fundamental and end the discussion.

Finally you agree. If I had said you can use built in types without including any headers while, for a class type you usually have to include a header, would you have accepted it as a valid difference to distinguish between built in types and class types? I don't think so. Likewise using std::is_fundamental as an argument would not have helped either. That is the reason why I mentioned that built in types are supported intrinsically by the compiler unlike the class types which would appear to be a more valid argument.

How do I create an interface for my app in c or c++

By interface you mean API, GUI or CLI?

GUI

GUI

Or at least an API for it to resemble a normal app in playstore like Kawlo or Photomath

If you want to write android app you can write GUI using native tools in Java and then connect the C++ logic to it (but it will probably require extra work if you want iOS also)

As of now I want to focus primarily on android apps thank a lot

https://developer.android.com/ndk/guides