Tensor Layout and Memory Access

2.2.1Shape, Stride, View, and Contiguous Tensor#

2.2.2Row-Major, Column-Major, and Blocked Layout#

2.2.3Coalescing, Vectorized Load/Store, and Alignment#

2.2.4Shared Memory Bank Conflict and Swizzle#

2.2.5Transpose, Packing, and Layout Conversion#

2.2.6Ragged Tensor, Padding, and Variable-Length Batches#