remove soulution :)

CUDA/add_Vec4.cu

@@ -33,9 +33,7 @@ int main(){
     cudaMemcpy(d_b, b, sizeof(float) * N, cudaMemcpyHostToDevice);
 
     // Executing kernel
-     int block_size = 512;
-    int grid_size = ((N + block_size) / block_size);
-    vector_add<<<grid_size,block_size>>>(d_out, d_a, d_b, N);
+    vector_add<<<1,256>>>(d_out, d_a, d_b, N);
 
     // Transfer data back to host memory
     cudaMemcpy(out, d_out, sizeof(float) * N, cudaMemcpyDeviceToHost);