#include <cuda.h>
#include <stdio.h>

__global__ void helloWorld(char* str){
  int idx = blockIdx.x * blockDim.x + threadIdx.x;
  str[idx] += idx;
}

int main(int argc, char** argv){
  char str[] = "Hello World";
  for(int i = 0; i < 12; i++){
    str[i] -= i;
  }

  char *d_str;
  size_t size = sizeof(str);
  cudaMalloc((void**)&d_str, size);
  cudaMemcpy(d_str, str, size, cudaMemcpyHostToDevice);

  dim3 dimGrid(2);
  dim3 dimBlock(6);

  helloWorld<<< dimGrid, dimBlock >>>(d_str);
  cudaMemcpy(str, d_str, size, cudaMemcpyDeviceToHost);
  cudaFree(d_str);

  printf("%s\\n", str);

  return 0;
}