pub fn dim_order(shape: &[usize], strides: &[usize]) -> Option<DimOrder>Expand description
The dimension order of a tensor that is dense in memory, or None if it is
not dense.
Dense means the strides are exactly a permutation of contiguous strides: no gaps, no overlap, no broadcasting. Use nested_dim_order when only an iteration order is needed and gaps in storage are acceptable.
Dimensions of size one are ignored while checking density — their stride is
arbitrary and carries no traffic — but they keep a position in the returned
order so it stays a permutation of 0..rank.