dpnp.ndarray.strides

property

property ndarray.strides

Tuple of bytes to step in each dimension when traversing an array.

The byte offset of element (i[0], i[1], ..., i[n]) in an array a is:

offset = sum(dpnp.array(i) * a.strides)

For full documentation refer to numpy.ndarray.strides.

See also

dpnp.lib.stride_tricks.as_strided

Return a view into the array with given shape and strides.

Examples

>>> import dpnp as np
>>> y = np.reshape(np.arange(2 * 3 * 4, dtype=np.int32), (2, 3, 4))
>>> y
array([[[ 0,  1,  2,  3],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]],
       [[12, 13, 14, 15],
        [16, 17, 18, 19],
        [20, 21, 22, 23]]], dtype=np.int32)
>>> y.strides
(48, 16, 4)
>>> y[1, 1, 1]
array(17, dtype=int32)
>>> offset = sum(i * s for i, s in zip((1, 1, 1), y.strides))
>>> offset // y.itemsize
17

>>> x = np.reshape(np.arange(5*6*7*8, dtype=np.int32), (5, 6, 7, 8))
>>> x = x.transpose(2, 3, 1, 0)
>>> x.strides
(32, 4, 224, 1344)
>>> offset = sum(i * s for i, s in zip((3, 5, 2, 2), x.strides))
>>> x[3, 5, 2, 2]
array(813, dtype=int32)
>>> offset // x.itemsize
813