#ifndef RESIDUAL_KERNEL_CUH_
#define RESIDUAL_KERNEL_CUH_

#include <cuda_runtime.h>
#include "../utils/cuda_utils.cuh"

__global__ void residual_forward_kernel(float* out, float* inp1, float* inp2, int N) {
    // Implement this
    int x = blockDim.x * blockIdx.x + threadIdx.x;

    if(x < N){
        out[x] = inp1[x] + inp2[x];
    }
}

// Launch kernel here
void residual_forward(float* out, float* inp1, float* inp2, int N) {
    // Implement this
    dim3 block(32,1,1);
    dim3 grid((N + block.x - 1)/ block.x);

    residual_forward_kernel<<<grid, block>>>(out, inp1, inp2, N);
}

#endif // RESIDUAL_KERNEL_CUH_