V0: vector addition on a single tile (1x1 grid of 1x1 tile group)
V1: vector addition on a single 1D tile group (1x1 grid of 2x1 tile group)
V2: vector addition on a single 2D tile group (1x1 grid of 4x4 tile group)
V3: vector addition on a 1D grid of 2D tile groups (4x1 grid of 2x2 tile groups)
The kernels for each version differs in index calculation, and it's targeted to teach about uses of tile group dimension and grid dimension, and how to distribute work among tiles/tile groups.
There are four versions to this code:
The kernels for each version differs in index calculation, and it's targeted to teach about uses of tile group dimension and grid dimension, and how to distribute work among tiles/tile groups.