CUDA Programming

Shared Memory

Fast, small-sized memory shared by all threads of the same block, enabling efficient communication between threads. Shared memory is much faster than global memory but limited in size per block.

← Indietro