#include <cuda_runtime.h>
#include <cutil_inline.h>
#include "fast_kernels.cu"

int main(int argc, char **argv){
  cudaSetDevice( 0 );
  short *h_in, *h_out, *d_out ;
  int size, bsx=16, bsy=16 ; ; 
  dim3 dimBlock, dimGrid ;
  cudaChannelFormatDesc channelDesc = cudaCreateChannelDesc<short>();
  cudaArray * array_img_in ;

  /* ... load image in h_in and determine H and W here ...*/
  cudaHostAlloc((void**)&h_out, size, cudaHostAllocDefault) ; 
  cudaMalloc((void**) &d_out, size);	  	  
  cudaMallocArray( &array_img_in, &channelDesc, W, H ); 
  cudaBindTextureToArray( tex_img_in, array_img_in, channelDesc);	  
  cudaMemcpyToArray( array_img_in, 0, 0, h_in, size, cudaMemcpyHostToDevice);
  dimBlock = dim3(bsx,bsy,1) ;
  dimGrid = dim3( L/dimBlock.x, H/dimBlock.y, 1) ;
  
  kernel_test<<< dimGrid, dimBlock, 0>>>(d_out, W, H) ;	
 
  cutilSafeCall( cudaMemcpy(h_out , d_out, size, cudaMemcpyDeviceToHost) ) ;
  
  /* ... use results here ... */
  cudaFreeHost(h_out) ;
  cudaFree(d_out);
  
  return 0;
}