CUDA shared memory, global memory, coalescing, L1 cache, memory hierarchy
Press ? anytime to show this help