13016.CUDA与C++混合编程demo

1 一个简单的c++与cuda编程示例

1.1 目录结构

1.2 CMakeLists.txt

cmake_minimum_required(VERSION 3.8)
project(cuda_demo VERSION 0.1)

# Find CUDA
find_package(CUDA REQUIRED)

# Add executable with source files
CUDA_ADD_EXECUTABLE(cuda_demo
    src/main.cpp
    src/cuda_test.cu
)

target_include_directories(cuda_demo PRIVATE /usr/local/cuda/include)
# Link CUDA libraries
target_link_libraries(cuda_demo ${CUDA_LIBRARIES})

1.3 cuda文件

// cuda_test.cu
#include <iostream>
#include <cuda_runtime.h>

__global__ void add(int *a, int *b, int *c, int N) {
    int idx = blockIdx.x * blockDim.x + threadIdx.x;
    if (idx < N) {
        c[idx] = a[idx] + b[idx];
    }
}

extern "C" void call_cuda_add(int *a, int *b, int *c, int N) {
    int blockSize = 256;
    int numBlocks = (N + blockSize - 1) / blockSize;

    add<<<numBlocks, blockSize>>>(a, b, c, N);
    cudaDeviceSynchronize();
}

1.4 main.cpp 文件

// main.cpp
#include <iostream>
#include <vector>
#include <cuda_runtime.h>

extern "C" void call_cuda_add(int *a, int *b, int *c, int N);

int main() {
    const int N = 1000;
    std::vector<int> a(N, 1), b(N, 2), c(N);

    int *d_a, *d_b, *d_c;
    cudaMalloc((void**)&d_a, N * sizeof(int));
    cudaMalloc((void**)&d_b, N * sizeof(int));
    cudaMalloc((void**)&d_c, N * sizeof(int));

    cudaMemcpy(d_a, a.data(), N * sizeof(int), cudaMemcpyHostToDevice);
    cudaMemcpy(d_b, b.data(), N * sizeof(int), cudaMemcpyHostToDevice);

    call_cuda_add(d_a, d_b, d_c, N);

    cudaMemcpy(c.data(), d_c, N * sizeof(int), cudaMemcpyDeviceToHost);

    for (int i = 0; i < N; ++i) {
        std::cout << "Result[" << i << "]: " << c[i] << std::endl;
    }

    cudaFree(d_a);
    cudaFree(d_b);
    cudaFree(d_c);

    return 0;
}

1.5 编译执行