The memory space on a GPU computation core which is shared by all threads of a block in a work grid (work-items in a work-group of the grid in OpenCL parlance).
