cuda编程[1]：一二三维网格和块的核函数

于 2024-09-03 15:53:30 发布
阅读量703 收藏 12
点赞数 10
分类专栏： cuda编程指南---官方英文文档版 文章标签： vscode c++ 机器学习 人工智能 深度学习
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/buuliuda/article/details/141863094
版权

目录

前言

所有的代码下载链接：code。以下代码展示了如何在 CUDA 中打印网格和线程的索引信息。代码包括一维、二维和三维的网格和块的设置，并定义了多个内核函数来输出当前的索引信息。

核函数

  1. 打印线程索引
__global__ void print_idx_kernel(){
    printf("block idx: (%3d, %3d, %3d), thread idx: (%3d, %3d, %3d)\n",
         blockIdx.z, blockIdx.y, blockIdx.x,
         threadIdx.z, threadIdx.y, threadIdx.x);
}
  1. 打印网格和块的维度
__global__ void print_dim_kernel(){
    printf("grid dimension: (%3d, %3d, %3d), block dimension: (%3d, %3d, %3d)\n",
         gridDim.z, gridDim.y, gridDim.x,
         blockDim.z, blockDim.y, blockDim.x);
}
  1. 打印每个块的线程索引
__global__ void print_thread_idx_per_block_kernel(){
    int index = threadIdx.z * blockDim.x * blockDim.y + \
              threadIdx.y * blockDim.x + \
              threadIdx.x;

    printf("block idx: (%3d, %3d, %3d), thread idx: %3d\n",
         blockIdx.z, blockIdx.y, blockIdx.x,
         index);
}
  1. 打印网格和块的维度
__global__ void print_thread_idx_per_grid_kernel(){
    int bSize  = blockDim.z * blockDim.y * blockDim.x;

    int bIndex = blockIdx.z * gridDim.x * gridDim.y + \
               blockIdx.y * gridDim.x + \
               blockIdx.x;

    int tIndex = threadIdx.z * blockDim.x * blockDim.y + \
               threadIdx.y * blockDim.x + \
               threadIdx.x;

    int index  = bIndex * bSize + tIndex;

    printf("block idx: %3d, thread idx in block: %3d, thread idx: %3d\n", 
         bIndex, tIndex, index);
}
  1. 打印坐标
__global__ void print_cord_kernel(){
    int index = threadIdx.z * blockDim.x * blockDim.y + \
              threadIdx.y * blockDim.x + \
              threadIdx.x;

    int x  = blockIdx.x * blockDim.x + threadIdx.x;
    int y  = blockIdx.y * blockDim.y + threadIdx.y;
    int z  = blockIdx.z * blockDim.z + threadIdx.z;

    printf("block idx: (%3d, %3d, %3d), thread idx: %3d, cord: (%3d, %3d, %3d)\n",
         blockIdx.z, blockIdx.y, blockIdx.x,
         index, x, y, z);
}

一维

不可以缺少，cudaDeviceSynchronize（ CPU与GPU端完成同步），当主函数在cpu中执行到需要调用核函数的时候不会等GPU全部完成返回结果，需要加上这个同步函数，不然运行可执行文件的时候会得到空的结果。

代码

void print_one_dim() {
    int inputSize = 8;
    int blockDim = 4;
    int gridDim = inputSize / blockDim;

    dim3 block(blockDim);
    dim3 grid(gridDim);

    printf("grid dimension: %d, block dimension: %d,\n", grid.x, block.x);

    cudaDeviceSynchronize();
}

二维

代码

void print_two_dim() {
    int inputWidth = 4;
    int blockDim = 2;
    int gridDim = inputWidth / blockDim;

    dim3 block(blockDim, blockDim);
    dim3 grid(gridDim, gridDim);

    printf("grid dimension: (%d, %d), block dimension: (%d, %d)\n",
           grid.y, grid.x, block.y, block.x);

    cudaDeviceSynchronize();
}

三维打印

代码

void print_three_dim() {
    int depth = 3;
    int height = 3;
    int width = 3;

    int blockDim = 2;

    dim3 block(blockDim, blockDim, blockDim);
    dim3 grid((width + blockDim - 1) / blockDim, 
              (height + blockDim - 1) / blockDim,
              (depth + blockDim - 1) / blockDim);

    printf("grid dimension: (%d, %d, %d), block dimension: (%d, %d, %d)\n",
           grid.z, grid.y, grid.x,
           block.z, block.y, block.x);

    cudaDeviceSynchronize();
}

主函数

这里就可以自己来使用print_one_dim, print_two_dim, print_three_dim测试不同网格不块的情况。可以自行组合定义核函数来测试所有情况。

int main() {
    // 选择打印的维度
    // print_one_dim();
    // print_two_dim();
    print_three_dim();

    return 0;
}

结果分析

这个只是一个小的.cu代码，所以我直接使用了笔记（点击代码链接可以看到）中得第一种方法编译。
在这里插入图片描述
打开当前代码目录下运行

nvcc grid_block_123D.cu -o test
./test

得到结果
···txt
grid dimension: (2, 2, 2), block dimension: (2, 2, 2)
block idx: ( 1, 0, 1), thread idx: 0, cord: ( 2, 0, 2)
block idx: ( 1, 0, 1), thread idx: 1, cord: ( 3, 0, 2)
block idx: ( 1, 0, 1), thread idx: 2, cord: ( 2, 1, 2)
block idx: ( 1, 0, 1), thread idx: 3, cord: ( 3, 1, 2)
block idx: ( 1, 0, 1), thread idx: 4, cord: ( 2, 0, 3)
block idx: ( 1, 0, 1), thread idx: 5, cord: ( 3, 0, 3)
block idx: ( 1, 0, 1), thread idx: 6, cord: ( 2, 1, 3)
block idx: ( 1, 0, 1), thread idx: 7, cord: ( 3, 1, 3)
block idx: ( 0, 1, 0), thread idx: 0, cord: ( 0, 2, 0)
block idx: ( 0, 1, 0), thread idx: 1, cord: ( 1, 2, 0)
block idx: ( 0, 1, 0), thread idx: 2, cord: ( 0, 3, 0)
block idx: ( 0, 1, 0), thread idx: 3, cord: ( 1, 3, 0)
block idx: ( 0, 1, 0), thread idx: 4, cord: ( 0, 2, 1)
block idx: ( 0, 1, 0), thread idx: 5, cord: ( 1, 2, 1)
block idx: ( 0, 1, 0), thread idx: 6, cord: ( 0, 3, 1)
block idx: ( 0, 1, 0), thread idx: 7, cord: ( 1, 3, 1)
block idx: ( 1, 0, 0), thread idx: 0, cord: ( 0, 0, 2)
block idx: ( 1, 0, 0), thread idx: 1, cord: ( 1, 0, 2)
block idx: ( 1, 0, 0), thread idx: 2, cord: ( 0, 1, 2)
block idx: ( 1, 0, 0), thread idx: 3, cord: ( 1, 1, 2)
block idx: ( 1, 0, 0), thread idx: 4, cord: ( 0, 0, 3)
block idx: ( 1, 0, 0), thread idx: 5, cord: ( 1, 0, 3)
block idx: ( 1, 0, 0), thread idx: 6, cord: ( 0, 1, 3)
block idx: ( 1, 0, 0), thread idx: 7, cord: ( 1, 1, 3)
block idx: ( 0, 0, 1), thread idx: 0, cord: ( 2, 0, 0)
block idx: ( 0, 0, 1), thread idx: 1, cord: ( 3, 0, 0)
block idx: ( 0, 0, 1), thread idx: 2, cord: ( 2, 1, 0)
block idx: ( 0, 0, 1), thread idx: 3, cord: ( 3, 1, 0)
block idx: ( 0, 0, 1), thread idx: 4, cord: ( 2, 0, 1)
block idx: ( 0, 0, 1), thread idx: 5, cord: ( 3, 0, 1)
block idx: ( 0, 0, 1), thread idx: 6, cord: ( 2, 1, 1)
block idx: ( 0, 0, 1), thread idx: 7, cord: ( 3, 1, 1)
block idx: ( 1, 1, 1), thread idx: 0, cord: ( 2, 2, 2)
block idx: ( 1, 1, 1), thread idx: 1, cord: ( 3, 2, 2)
block idx: ( 1, 1, 1), thread idx: 2, cord: ( 2, 3, 2)
block idx: ( 1, 1, 1), thread idx: 3, cord: ( 3, 3, 2)
block idx: ( 1, 1, 1), thread idx: 4, cord: ( 2, 2, 3)
block idx: ( 1, 1, 1), thread idx: 5, cord: ( 3, 2, 3)
block idx: ( 1, 1, 1), thread idx: 6, cord: ( 2, 3, 3)
block idx: ( 1, 1, 1), thread idx: 7, cord: ( 3, 3, 3)
block idx: ( 0, 1, 1), thread idx: 0, cord: ( 2, 2, 0)
block idx: ( 0, 1, 1), thread idx: 1, cord: ( 3, 2, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 7, cord: ( 3, 3, 1)
block idx: ( 0, 0, 0), thread idx: 0, cord: ( 0, 0, 0)
block idx: ( 0, 0, 0), thread idx: 1, cord: ( 1, 0, 0)
block idx: ( 0, 0, 0), thread idx: 2, cord: ( 0, 1, 0)
block idx: ( 0, 0, 0), thread idx: 3, cord: ( 1, 1, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 7, cord: ( 3, 3, 1)
block idx: ( 0, 0, 0), thread idx: 0, cord: ( 0, 0, 0)
block idx: ( 0, 0, 0), thread idx: 1, cord: ( 1, 0, 0)
block idx: ( 0, 0, 0), thread idx: 2, cord: ( 0, 1, 0)
block idx: ( 0, 0, 0), thread idx: 3, cord: ( 1, 1, 0)
block idx: ( 0, 0, 0), thread idx: 4, cord: ( 0, 0, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 7, cord: ( 3, 3, 1)
block idx: ( 0, 0, 0), thread idx: 0, cord: ( 0, 0, 0)
block idx: ( 0, 0, 0), thread idx: 1, cord: ( 1, 0, 0)
block idx: ( 0, 0, 0), thread idx: 2, cord: ( 0, 1, 0)
block idx: ( 0, 0, 0), thread idx: 3, cord: ( 1, 1, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 7, cord: ( 3, 3, 1)
block idx: ( 0, 0, 0), thread idx: 0, cord: ( 0, 0, 0)
block idx: ( 0, 0, 0), thread idx: 1, cord: ( 1, 0, 0)
block idx: ( 0, 0, 0), thread idx: 2, cord: ( 0, 1, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 7, cord: ( 3, 3, 1)
block idx: ( 0, 0, 0), thread idx: 0, cord: ( 0, 0, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 7, cord: ( 3, 3, 1)
block idx: ( 0, 0, 0), thread idx: 0, cord: ( 0, 0, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 7, cord: ( 3, 3, 1)
block idx: ( 0, 0, 0), thread idx: 0, cord: ( 0, 0, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 2, cord: ( 2, 3, 0)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 3, cord: ( 3, 3, 0)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 4, cord: ( 2, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 5, cord: ( 3, 2, 1)
block idx: ( 0, 1, 1), thread idx: 6, cord: ( 2, 3, 1)
block idx: ( 0, 1, 1), thread idx: 7, cord: ( 3, 3, 1)
block idx: ( 0, 0, 0), thread idx: 0, cord: ( 0, 0, 0)
block idx: ( 0, 0, 0), thread idx: 1, cord: ( 1, 0, 0)
block idx: ( 0, 0, 0), thread idx: 2, cord: ( 0, 1, 0)
block idx: ( 0, 0, 0), thread idx: 3, cord: ( 1, 1, 0)
block idx: ( 0, 0, 0), thread idx: 4, cord: ( 0, 0, 1)
block idx: ( 0, 0, 0), thread idx: 5, cord: ( 1, 0, 1)
block idx: ( 0, 0, 0), thread idx: 6, cord: ( 0, 1, 1)
block idx: ( 0, 0, 0), thread idx: 7, cord: ( 1, 1, 1)
block idx: ( 1, 1, 0), thread idx: 0, cord: ( 0, 2, 2)
block idx: ( 1, 1, 0), thread idx: 1, cord: ( 1, 2, 2)
block idx: ( 1, 1, 0), thread idx: 2, cord: ( 0, 3, 2)
block idx: ( 1, 1, 0), thread idx: 3, cord: ( 1, 3, 2)
block idx: ( 1, 1, 0), thread idx: 4, cord: ( 0, 2, 3)
block idx: ( 1, 1, 0), thread idx: 5, cord: ( 1, 2, 3)
block idx: ( 1, 1, 0), thread idx: 6, cord: ( 0, 3, 3)
block idx: ( 1, 1, 0), thread idx: 7, cord: ( 1, 3, 3)

小马敲马
关注
  • 10
    点赞
  • 踩
  • 12
    收藏
    觉得还不错? 一键收藏
  • 打赏
    打赏
  • 0
    评论
专栏目录
3维线程格 gpu_GPU编程2--CUDA核函数和线程配置
weixin_39575502的博客
12-23 280
CUDA核函数在GPU上执行的函数称为CUDA核函数(Kernel Function)，核函数会被GPU上多个线程执行，我们可以在核函数中获取当前线程的ID。//CUDA核函数的定义__global__voidaddKernel(int*c,constint*a,constint*b){inti=threadIdx.x;c[i]=a[i]+b[i];}//CUD...
CUDA.jl：Julia中的CUDA编程
02-03
之后，你可以通过指定线程块大小和网格大小来启动这个函数。例如： ```julia using CUDA @cuda function add!(a, b, c) i = (blockIdx().x - 1) * blockDim().x + threadIdx().x if i (a) c[i] = a[i] + b[i] ...
CUDA编程：其二、网格与线程块
jiexianxiaogege的博客
04-16 472
前文提到了GPU的强大之处是处理并行，那自然是要开大量线程。本文主要描述GPU如何组织大量线程。
CUDA C编程10:核函数可达到的带宽
DU_YULIN的博客
04-19 1348
系列文章目录 文章目录系列文章目录前言一、理论知识二、案例分享2.1 朴素转置2.2 展开转置2.3 对角转置2.4 通过瘦块方法来增加并行性2.5 完整代码：总结参考资料 前言 忙里偷闲，继续学习CUDA C编程，今天开始学习核函数的带宽的相关知识点，提高性能。 一、理论知识 分析核函数性能时，要注意以下两点： （1）内存延迟：完成一次内存请求的时间； （2）内存带宽：SM访问设备内存的速度，以每单位时间内的字节数来度量。 之前了解的改进核函数性能的方法有两种： （1）通过最大化并行线程束
CUDA编程：线程模型
qq_43570025的博客
01-03 692
cuda线程模型
CUDA并行编程基础：线程、块和网格
# 1. 介绍 ## 1.1 CUDA并行编程概述 CUDA（Compute Unified Device Architecture）是由NVIDIA推出的用于通用目的...CUDA架构包括多个核心概念，包括线程、线程块和网格。在接下来的章节中，我们将详细介绍这些概念并
cuda_practice:Cuda C ++编程指南
04-06
1. **CUDA架构**：了解GPU的多核心架构，包括流处理器（Streaming Multiprocessors, SMs）、线程块（Thread Blocks）、线程网格（Thread Grids）以及共享内存和全局内存的概念。 2. **CUDA编程模型**：学习如何定义...
cuda-experiments:CUDA实验
04-10
4. **CUDA编程接口**：CUDA的核心API包括`cudaMalloc`、`cudaMemcpy`、`cudaFree`、`cudaLaunchKernel`等函数，用于管理GPU内存、数据传输和启动计算kernel。 5. **Kernel函数**：Kernel是CUDA程序中的核心部分，它...
vp9_cuda_encoder:使用CUDA并行编程使vp9编码器加速
05-10
2. **线程块与网格**：CUDA编程的核心是线程和线程块的组织。线程块可以并行运行在GPU的不同流处理器上，线程块之间又可以组成网格。对于VP9编码，可以根据编码任务的特性，设计合适的线程组织结构，例如，每个线程...
vscode使用
Q_004的博客
08-30 1133
备注：有人使用"vue",我开始也试了vue，不知道为啥不可以，改成html就可以了。点击文件-》首选项-》设置 或者直接快捷键ctrl+，在搜索框中输入settings.json,打开此文件。重新打开vscode就可以i了。添加对.vue文件的配置。
HPM5301系列--VSCODE开发环境问题修复(一)
coderma的博客
08-27 903
解决CMAKE_BUILD_TYPE值大小写问题
Cubemx生成的MDK文件，用vscode的EIED插件开发，会找不到头文件
2303_76872898的博客
08-28 285
避坑，在生成文件的时候，不要选生成所有的文件。选第二个，就可以解决问题了，成功编译。
vscode添加到环境变量之快捷使用
m0_56741459的博客
08-29 548
将 VSCode 添加到环境变量中，主要是为了方便在命令行中全局调用code命令，从而更高效地打开和编辑文件或项目，提升开发效率，尤其是在命令行和编辑器结合的工作流中。
如何在VSCode中同时打开多个页面？
seasonsyy的博客
09-02 433
如何在VSCode中同时打开多个页面？
c++命令模式
qq_71286244的博客
08-30 4074
c++命令模式详细介绍
c语言赋值截断
hutaotaotao的博客
09-02 417
c语言赋值过程与截断行为
【C++】智能指针——auto_ptr，unique_ptr，shared_ptr
2301_79796701的博客
09-01 1196
1.智能指针对象中引用计数是多个智能指针对象共享的，两个线程中智能指针的引用计数同时 ++或--，这个操作不是原子的，引用计数原来是1，++了两次，可能还是2.这样引用计数就错 乱了。旧的智能指针的会转移给新的智能指针，而旧的智能指针的会被置空。解决方案：在引用计数的场景下，把节点中的_prev和_next改成weak_ptr就可以了，weak_ptr不会增加引用计数，代码示意。该问题是有类似于链表节点的结构引起的，当外部的智能指针都释放了，就拿不内部的智能指针，那么节点就不会释放。使用方法参考官方文档。
C++友类匿名对象，运算符重载
最新发布
m0_75109289的博客
09-02 115
【代码】C++友类匿名对象，运算符重载。
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包

打赏作者

小马敲马

你的鼓励将是我创作的最大动力

¥1 ¥2 ¥4 ¥6 ¥10 ¥20
扫码支付：¥1
获取中
扫码支付

您的余额不足，请更换扫码支付或充值

打赏作者

实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值