(1 pt) With matrix size 1024x1024, how much shared memory (in bytes) is allocated in each block in the mm-cuda2 version? To earn full credits, you need to show your calculations manually and use Nvidia profiling to verify the results. Attach a screenshot of profiling results