Hi everyone, I’m just wondering if this memcpy is allowed. In particular, the fact that I’m using an offset into the array.
float* host_array = (float*) malloc(size);
float* device_array;
cudaMalloc( (void**) &device_array, size*size));
cudaMemcpy(host_array, device_array + offset, size, cudaMemcpyDeviceToHost);
Thanks!
I think you can try it by yourself, but anyway, the answer is yes. Good luck!