:mod:`heat.memory`
=======================
.. py:module:: heat.core.memory

.. autoapi-nested-parse::

   Utilities to manage the internal memory of an array.



Module Contents
---------------


.. function:: copy(x: heat.core.dndarray.DNDarray) -> heat.core.dndarray.DNDarray

   Return a deep copy of the given object.

   :param x: Input array to be copied.
   :type x: DNDarray

   .. rubric:: Examples

   >>> a = ht.array([1, 2, 3])
   >>> b = ht.copy(a)
   >>> b
   DNDarray([1, 2, 3], dtype=ht.int64, device=cpu:0, split=None)
   >>> a[0] = 4
   >>> a
   DNDarray([4, 2, 3], dtype=ht.int64, device=cpu:0, split=None)
   >>> b
   DNDarray([1, 2, 3], dtype=ht.int64, device=cpu:0, split=None)


.. function:: sanitize_memory_layout(x: torch.Tensor, order: str = 'C') -> torch.Tensor

   Return the given object with memory layout as defined below. The default memory distribution is assumed.

   :param x: Input data
   :type x: torch.Tensor
   :param order: Default is ``'C'`` as in C-like (row-major) memory layout. The array is stored first dimension first (rows first if ``ndim=2``).
                 Alternative is ``'F'``, as in Fortran-like (column-major) memory layout. The array is stored last dimension first (columns first if ``ndim=2``).
   :type order: str, optional.