1
#include <cuda_runtime.h>
2

3
__global__ void matrix_add(const float* A, const float* B, float* C, int N) {
4
    int i = blockIdx.x * blockDim.x + threadIdx.x;
5
    if (i < N * N) {
6
        C[i] = A[i] + B[i];
7
    }
8
}
9

10
// A, B, C are device pointers (i.e. pointers to memory on the GPU)
11
extern "C" void solve(const float* A, const float* B, float* C, int N) {
12
    int threadsPerBlock = 256;
13
    int blocksPerGrid = (N * N + threadsPerBlock - 1) / threadsPerBlock;
14

15
    matrix_add<<<blocksPerGrid, threadsPerBlock>>>(A, B, C, N);
16
    cudaDeviceSynchronize();
17
}