René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"`dpt.astype` behaves differently on Windows depending on type of input array","articleBody":"There is a suspicious behavior I faced on Windows (LNL with fp64 support for GPU):\n```python\nimport dpctl, dpctl.tensor as dpt, numpy as np\n\ndpctl.__version__\n# Out: '0.22.0dev0+117.g7960513f68'\n\n# for integer input the result is expected\na = dpt.asarray([160, 120])\ndpt.astype(a, dpt.uint8)\n# Out: usm_ndarray([160, 120], dtype=uint8)\n\n# for float32 the result seems incorrect:\na = dpt.asarray([160., 120.], dtype='f4')\ndpt.astype(a, dpt.uint8)\n# Out: usm_ndarray([127, 120], dtype=uint8)\n\n# but with float64 works as expected:\na = dpt.asarray([160., 120.], dtype='f8')\ndpt.astype(a, dpt.uint8)\n# Out: usm_ndarray([160, 120], dtype=uint8)\n\n# NumPy returns the same result in both cases:\n\na = np.asarray([160, 120])\nnp.astype(a, dpt.uint8)\n# Out: array([160, 120], dtype=uint8)\n\na = np.asarray([160., 120.], dtype='f4')\nnp.astype(a, dpt.uint8)\n# Out: array([160, 120], dtype=uint8)\n\n# for CPU device the result is the same as for NumPy:\n\na = dpt.asarray([160., 120.], dtype='i4', device='cpu')\ndpt.astype(a, dpt.uint8)\n# Out: usm_ndarray([160, 120], dtype=uint8)\n```\n\nI wonder if there is any limitation for float32 dtype or an issue.\n","author":{"url":"https://github.com/antonwolfy","@type":"Person","name":"antonwolfy"},"datePublished":"2026-01-15T14:41:17.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":1},"url":"https://github.com/2230/dpctl/issues/2230"}