compile error fixex

mean_shift_cuda/meanshift_kernels.cu

@@ -54,8 +54,9 @@ __global__ void denominator_kernel<<<dimGrid, dimBlock>>>(Matrix denominator, Ma
     int col = blockIdx.y * blockDim.y + threadIdx.y;
 
 
-    if (row>=total || col>=total)
+    if (row * denominator.width + col > denominator.width * denominator.height){
         return;
+    }
 
     denominator[col]=0;
     denominator[row] += kernel_matrix[row*denominator.width + col];

mean_shift_cuda/meanshift_kernels.h

@@ -13,6 +13,6 @@ __global__ void multiply_kernel(Matrix matrix1, Matrix matrix2, Matrix output);
 __global__ void calculate_kernel_matrix_kernel(Matrix shifted_points, Matrix original_points
     , double deviation, Matrix kernel_matrix);
 
-__global__ void denominator_kernel(Matrix denominator, Matrix kernel_matrix, int T);
+__global__ void denominator_kernel(Matrix denominator, Matrix kernel_matrix);
 
 #endif //SERIAL_KERNELS_H

mean_shift_cuda/meanshift_utils.cu

@@ -197,7 +197,7 @@ int meanshift(double **original_points, double ***shifted_points, int deviation
 //        }
 //        denominator[i] = sum;
 //    }
-    calculate_denominator(kernel_matrix);
+    denominator = calculate_denominator(kernel_matrix);
 
     // creates new y vector
     double **new_shift = alloc_2d_double(NUMBER_OF_POINTS, DIMENSIONS);
@@ -469,7 +469,7 @@ double * calculate_denominator(double **kernel_matrix){
         first_iter = false;
     }
 
-    denominator_kernel<<<dimGrid, dimBlock>>>(d_denominator_matrix, d_kernel_matrix, T);
+    denominator_kernel<<<dimGrid, dimBlock>>>(d_denominator_matrix, d_kernel_matrix);
     gpuErrchk( cudaPeekAtLastError() );
     gpuErrchk( cudaDeviceSynchronize() );
 

mean_shift_cuda/meanshift_utils.h

@@ -73,6 +73,6 @@ void save_matrix(double **matrix
 
 //Function calculate_denominator allocates memory in GPU, sends the data and calls the
 //denominator kernel function.
-void calculate_denominator(double **kernel_matrix);
+double * calculate_denominator(double **kernel_matrix);
 
 #endif //SERIAL_UTILS_H