dpctl

dpctl copied to clipboard

Review of contributor guides

1

- [ ] Have you provided a meaningful PR description? - [ ] Have you added a test, reproducer or referred to an issue with a reproducer? - [ ]...

bdmoore1

Review of User Guides section

4

- [ ] Have you provided a meaningful PR description? - [ ] Have you added a test, reproducer or referred to an issue with a reproducer? - [ ]...

bdmoore1

The below example demonstrates different behavior in `reshape` method of NumPy and dpctl: ```python import numpy, dpctl, dpctl.tensor as dpt dpctl.__version__ # Out: '0.17.0dev0+325.g0cb2181547' a = dpt.ones((2, 1)) a.shape, a.strides...

antonwolfy

Meta data improvements

12

Changes in this PR: - Move all supported configuration to `pyproject.toml` from `setup.py` and `setup.cfg` (latest one was removed). - Populate additional fields to `project` and `build-system` sections of `pyproject.toml`....

ZzEeKkAa

less_equal returns incorrect result

2

The code below works incorrectly in dpctl: ``` import numpy, dpctl, dpctl.tensor as dpt dpt.less_equal(dpt.asarray(2, dtype=numpy.int32), numpy.iinfo(numpy.uint32).max) Out: usm_ndarray(False) numpy.less_equal(numpy.asarray(2, dtype=numpy.int32), numpy.iinfo(numpy.uint32).max) Out: True ``` It looks like dpctl always...

antonwolfy

Segmentation fault with dpctl.tensor.argsort on CPU device

1

The below example crashes with Segmentation fault while running on CPU device: ```python import dpctl.tensor as dpt from numpy.random import default_rng N = 4 * 10**6 rng = default_rng(42) a...

antonwolfy

`floor_divide` returns different result for arrays of floating dtype on GPU and CPU devices

5

In below example the behavior is different between CPU and GPU devices: ```python import numpy, dpctl, dpctl.tensor as dpt dpctl.__version__ # Out: '0.17.0dev0+300.g7757857466' a = dpt.arange(1, 10, dtype='f', device='gpu') b...

antonwolfy

Zeroed result returned by an elementwise function

4

Sometimes a small test below ```python import dpctl, dpctl.tensor as dpt import numpy list_of_backend_str = [ "host", "level_zero", "opencl", ] list_of_device_type_str = [ "host", "gpu", "cpu", ] available_devices = [...

antonwolfy

[DO NOT MERGE] Add hand-pinning of runtime

2

This PR is to test hand-added pinning for DPC++ runtime packages the `dpctl` is compatible with - [x] Have you provided a meaningful PR description? - [ ] Have you...

oleksandr-pavlyk

How to convert CUDA-Python to SYCL using dpctl?

1

Hi team, Would you kindly provide guidance on transitioning CUDA-Python code to SYCL? Specifically, I am interested in converting the code available [here](https://github.com/NVIDIA/cuda-python/blob/main/examples/0_Introduction/vectorAddDrv_test.py) to SYCL utilizing dpctl. And also, is...

AryanKarumuri