dpctl.tensor.concat() failed on Iris Xe

[antonwolfy]

The below example triggers an exception on Iris Xe, but expected to be working:

import numpy, dpctl, dpctl.tensor as dpt

a = dpt.arange(10, dtype=numpy.int64)
b = dpt.arange(10, dtype=numpy.float32)

dpt.concat((a, b))
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[4], line 1
----> 1 dpt.concat((a, b))

File ~/miniconda3/envs/dpnp_dev/lib/python3.9/site-packages/dpctl/tensor/_manipulation_functions.py:615, in concat(arrays, axis)
    609     res_shape_axis = res_shape_axis + X.shape[axis]
    611 res_shape = tuple(
    612     X0_shape[i] if i != axis else res_shape_axis for i in range(X0.ndim)
    613 )
--> 615 res = dpt.empty(
    616     res_shape, dtype=res_dtype, usm_type=res_usm_type, sycl_queue=exec_q
    617 )
    619 hev_list = []
    620 fill_start = 0

File ~/miniconda3/envs/dpnp_dev/lib/python3.9/site-packages/dpctl/tensor/_ctors.py:739, in empty(shape, dtype, order, device, usm_type, sycl_queue)
    737 sycl_queue = normalize_queue_device(sycl_queue=sycl_queue, device=device)
    738 dtype = _get_dtype(dtype, sycl_queue)
--> 739 _ensure_native_dtype_device_support(dtype, sycl_queue.sycl_device)
    740 res = dpt.usm_ndarray(
    741     shape,
    742     dtype=dtype,
   (...)
    745     buffer_ctor_kwargs={"queue": sycl_queue},
    746 )
    747 return res

File ~/miniconda3/envs/dpnp_dev/lib/python3.9/site-packages/dpctl/tensor/_ctors.py:313, in _ensure_native_dtype_device_support(dtype, dev)
    300 """Check that dtype is natively supported by device.
    301
    302 Arg:
   (...)
    310    ValueError - if device does not natively support this dtype.
    311 """
    312 if dtype in [dpt.float64, dpt.complex128] and not dev.has_aspect_fp64:
--> 313     raise ValueError(
    314         f"Device {dev.name} does not provide native support "
    315         "for double-precision floating point type."
    316     )
    317 if (
    318     dtype
    319     in [
   (...)
    322     and not dev.has_aspect_fp16
    323 ):
    324     raise ValueError(
    325         f"Device {dev.name} does not provide native support "
    326         "for half-precision floating point type."
    327     )

ValueError: Device Intel(R) Graphics [0x9a49] does not provide native support for double-precision floating point type.

dpctl.__version__
# Out: '0.14.6dev0+84.gbf2d1ff3a'