Return default floating point type of the device in dpctl.tensor.astype when newdtype=None (#1262)

* For None returns default floating point type supported by device.
* For all other types returns dpt.dtype(dtype)

---------

Co-authored-by: Oleksandr Pavlyk <oleksandr.pavlyk@intel.com>

Author: vlad-perevezentsev

dpctl/tensor/_copy_utils.py

@@ -23,6 +23,7 @@
 import dpctl.tensor as dpt
 import dpctl.tensor._tensor_impl as ti
 import dpctl.utils
+from dpctl.tensor._ctors import _get_dtype
 from dpctl.tensor._device import normalize_queue_device
 
 __doc__ = (
@@ -364,7 +365,8 @@ def astype(usm_ary, newdtype, order="K", casting="unsafe", copy=True):
         array (usm_ndarray):
             An input array.
         new_dtype (dtype):
-            The data type of the resulting array.
+            The data type of the resulting array. If `None`, gives default
+            floating point type supported by device where `array` is allocated.
         order ({"C", "F", "A", "K"}, optional):
             Controls memory layout of the resulting array if a copy
             is returned.
@@ -392,7 +394,7 @@ def astype(usm_ary, newdtype, order="K", casting="unsafe", copy=True):
             "Recognized values are 'A', 'C', 'F', or 'K'"
         )
     ary_dtype = usm_ary.dtype
-    target_dtype = dpt.dtype(newdtype)
+    target_dtype = _get_dtype(newdtype, usm_ary.sycl_queue)
     if not dpt.can_cast(ary_dtype, target_dtype, casting=casting):
         raise TypeError(
             f"Can not cast from {ary_dtype} to {newdtype} "

dpctl/tests/test_usm_ndarray_ctor.py

@@ -1196,6 +1196,11 @@ def test_astype():
     assert np.allclose(dpt.to_numpy(Y), np.full(Y.shape, 7, dtype="f4"))
     Y = dpt.astype(X[::2, ::-1], "i4", order="K", copy=False)
     assert Y.usm_data is X.usm_data
+    Y = dpt.astype(X, None, order="K")
+    if X.sycl_queue.sycl_device.has_aspect_fp64:
+        assert Y.dtype is dpt.float64
+    else:
+        assert Y.dtype is dpt.float32
 
 
 def test_astype_invalid_order():