A noinline inline function? What sorcery is this?

栏目: IT技术 · 发布时间: 4年前

内容简介:May 21st, 2020You can declare a noinline inline function.What sorcery is this, a function that is both inline and not-inline?
A noinline inline function? What sorcery is this?

Raymond

May 21st, 2020

You can declare a noinline inline function.

void g();

// gcc
__attribute__((noinline)) inline void f()
{
    g();
}

// MSVC
__declspec(noinline) inline void f()
{
    g();
}

void tryme()
{
    f();
    f();
}

What sorcery is this, a function that is both inline and not-inline?

The two keywords are not contradictory because they describe different senses of the word “inline”.

The C++ language keyword inline means “can be defined in multiple translation units without triggering an ODR violation.” In other words, it lets you put the function definition in a header file that is included by multiple C++ files.

The function attribute/declaration specifier noinline means “do not inline this function during code generation.” It is a directive to the optimizer not to perform inline substitution during code generation.

Historically, the inline C++ keyword was originally an optimizer hint, but optimizers were given permission to ignore it and make their own decisions about inline substitution during code generation. Nowadays, compilers pretty much ignore the optimization aspect of the inline keyword. The only thing that remains of the inline keyword is the ability to have multiple definitions without violating ODR.

You could say that the modern sense of the C++ keyword inline is “defined right here.” It’s a statement about the source code, not the object code.

In the above example, the function f is a noinline inline function. The inline keyword allows the definition of f to go into a header file that is consumed by multiple translation units. The noinline attribute/declaration specifier tells the optimizer to emit code for f and call it, rather than embedding the body of f into its call sites. The function tryme will call the function f twice, instead of optimizing out the call and just calling g twice.


以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持 码农网

查看所有标签

猜你喜欢:

本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们

数据结构与算法分析(C++版)(第3版)

数据结构与算法分析(C++版)(第3版)

Clifford A. Shaffer / 张铭、刘晓丹、等译 / 电子工业出版社 / 2013 / 59.00元

本书采用当前流行的面向对象的C++程序设计语言来描述数据结构和算法, 因为C++语言是程序员最广泛使用的语言。因此, 程序员可以把本书中的许多算法直接应用于将来的实际项目中。尽管数据结构和算法在设计本质上还是很底层的东西, 并不像大型软件工程项目开发那样, 对面向对象方法具有直接的依赖性, 因此有人会认为并不需要采用高层次的面向对象技术来描述底层算法。 但是采用C++语言能更好地体现抽象数据类型的......一起来看看 《数据结构与算法分析(C++版)(第3版)》 这本书的介绍吧!

在线进制转换器
在线进制转换器

各进制数互转换器

随机密码生成器
随机密码生成器

多种字符组合密码

SHA 加密
SHA 加密

SHA 加密工具