C++:静态断言内存对齐

静态断言

C中的断言assert
(1)直接参考：https://www.cnblogs.com/lvchaoshun/p/7816288.html
(2)C的assert是运行时检测发现错误，而不是编译时
(3)C在编译时错误用#error来输出
C++静态断言
(1)C++引入static_assert(表达式, “提示字符串”)来实现编译时的静态断言
(2)实例演示

#include <iostream>
#include <type_traits>// 编译期常量表达式
constexpr int factorial(int n) {return (n <= 1) ? 1 : (n * factorial(n - 1));
}int main() {// 编译时断言：检查某些条件是否成立static_assert(sizeof(int) == 4, "int 类型的大小不是 4 字节");static_assert(factorial(5) == 120, "Factorial 计算错误");std::cout << "所有静态断言均通过。" << std::endl;return 0;
}

#include <iostream>
#include <cassert>int factorial(int n) {assert(n >= 0); // 断言n是非负数if (n == 0) {return 1;} else {return n * factorial(n - 1);}
}int main() {int num1 = 5;int num2 = -3;// 正常情况std::cout << "Factorial of " << num1 << " is " << factorial(num1) << std::endl;// 异常情况（将触发断言）std::cout << "Factorial of " << num2 << " is " << factorial(num2) << std::endl;return 0;
}

静态断言主要用途
(1)static_assert主要用于检查模板参数是否符合期望
(2)C++20中引入了concept来进一步更好的实现模板参数的编译时类型匹配检查

内存对齐

C语言中内存对齐关键点
(1)#pragma 和 attribute((packed)) attribute((aligned(n)))

#include <stdio.h>// 使用 #pragma pack 控制对齐
#pragma pack(push, 1) // 设置为1字节对齐typedef struct {char c;int i;double d;
} PackedStruct;#pragma pack(pop) // 恢复默认对齐// 使用 __attribute__((packed)) 控制对齐
typedef struct {char c;int i;double d;
} __attribute__((packed)) PackedStructAttribute;// 使用 __attribute__((aligned(n))) 控制对齐
typedef struct {char c;int i;double d;
} AlignedStruct __attribute__((aligned(16)));int main() {PackedStruct ps;PackedStructAttribute psa;AlignedStruct as;printf("Size of PackedStruct: %zu bytes\n", sizeof(PackedStruct));printf("Size of PackedStructAttribute: %zu bytes\n", sizeof(PackedStructAttribute));printf("Size of AlignedStruct: %zu bytes\n", sizeof(AlignedStruct));printf("Address of ps.c: %p\n", (void*)&ps.c);printf("Address of ps.i: %p\n", (void*)&ps.i);printf("Address of ps.d: %p\n", (void*)&ps.d);printf("Address of psa.c: %p\n", (void*)&psa.c);printf("Address of psa.i: %p\n", (void*)&psa.i);printf("Address of psa.d: %p\n", (void*)&psa.d);printf("Address of as.c: %p\n", (void*)&as.c);printf("Address of as.i: %p\n", (void*)&as.i);printf("Address of as.d: %p\n", (void*)&as.d);return 0;
}

在这里插入图片描述

C++中内存对齐新增关键字
(1)alignof (C++11 起) 查询对齐要求
(2)alignas (C++11 起)设置对齐,效果：和__attribute__((aligned(n)))效果一样,往大了设置有用

#include <iostream>
#include <cstddef> // for std::size_t// 使用 alignas 设置对齐
struct alignas(16) AlignedStruct {char c;int i;double d;
};// 未对齐的结构体
struct UnalignedStruct {char c;int i;double d;
};int main() {// 使用 alignof 查询对齐要求std::cout << "Alignment of char: " << alignof(char) << std::endl;std::cout << "Alignment of int: " << alignof(int) << std::endl;std::cout << "Alignment of double: " << alignof(double) << std::endl;std::cout << "Alignment of AlignedStruct: " << alignof(AlignedStruct) << std::endl;std::cout << "Alignment of UnalignedStruct: " << alignof(UnalignedStruct) << std::endl;// 打印结构体大小std::cout << "Size of AlignedStruct: " << sizeof(AlignedStruct) << std::endl;std::cout << "Size of UnalignedStruct: " << sizeof(UnalignedStruct) << std::endl;// 分配对齐的内存void* ptr = aligned_alloc(alignof(AlignedStruct), sizeof(AlignedStruct));if (ptr) {std::cout << "Memory allocated at address: " << ptr << std::endl;free(ptr); // 记得释放内存} else {std::cerr << "Memory allocation failed!" << std::endl;}return 0;
}

在这里插入图片描述

什么情况下需要人为改变/指定对齐方式
(1)往大去对齐。有时候会有一些硬件特殊要求，譬如MMU，cache等。用__attribute__((aligned(n)))实测ok，用#pragma实测不ok
(2)往下去对齐。有时候需要节省内存而浪费效率，所以希望忽略内存对齐，紧密排放。用#pramgma实测ok，用__attribute__((aligned(n)))实测不ok