Interleave Arrays

1
#include <cuda_runtime.h>
2

3
__global__ void interleave_kernel(const float* A, const float* B, float* output, int N) {
4
    int idx = blockDim.x * blockIdx.x + threadIdx.x;
5
    if (idx >= N) return;
6
    output[2 * idx] = A[idx];
7
    output[2 * idx + 1] = B[idx];
8
}
9

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

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